Search

Wheat scab has been one of the most severe diseases for grains in recent years. Unmanned aerial vehicles (UAV) can be used to monitor the wheat scab using multispectral remote sensing. In this study, a series of operations were carried out to eliminate the inter-ridge background elements (such as soil and shadow intercage background elements) under natural conditions. The precision segmentation of inter-ridge background was then realized to improve the identification accuracy of wheat scab. A high-precision monitoring model was also constructed for the wheat scab. Specifically, the research object was selected as the wheat at the filling stage in the Zhenjiang Academy of Agricultural Sciences in Jiangsu Province, China. Rededge MX multispectral camera (five channels) was installed on a DJI M600 Pro six-rotor UAV to capture the multispectral images of the wheat canopy. Multiple vegetation indices were then calculated using the spectral features of the wheat canopy. A correlation analysis was implemented to determine the highest correlation vegetation indices (VIs) with the wheat scab, and the gray scale co-occurrence matrix (gray-level co-occurrence matrix, GLCM and texture feature (TFs). Otus, threshold segmentation and support vector machine (SVM) were used for the semantic segmentation of wheat scab multispectral images. There was a reduced influence of field edge shadow and soil background elements on the identification accuracy of wheat scab. Partial least squares regression (PLSR) was used after eliminating the ridge background. The regression model was established for the VIs, TFs, Vis&TFs and wheat scab disease index (DI). The regression equation of the monitoring model was then constructed to draw the change chart of disease severity. The experimental results show that the best performance was achieved in the visual interpretation threshold segmentation (overall accuracy: 92.06%, Kappa coefficient: 0.84), and the OTSU threshold segmentation (overall accuracy: 90.52%, Kappa coefficient: 0.81) in the inter-row background elimination model. In the wheat scab monitoring model, the accuracy of the Vis&TFs model was higher than that of the VIs and TFs model, and the accuracy of the refined semantic segmentation model was better than that of the unsegmentation model. Taking the model on May 8 as an example, the training set and validation set R2 of the VIs-PLSR undivided model were 0.71 and 0.73, RMSE were 5.61 and 6.72, and RPD were 1.83 and 1.89, respectively. The training set and validation set R2 of the TFs-PLSR undivided model were 0.64 and 0.62, RMSE were 6.03 and 6.92, and RPD were 1.67 and 1.69, respectively. Vis&TFs undivided model training set and validation set R2 were 0.73 and 0.61, RMSE were 6.31 and 6.68, RPD were 1.93 and 1.89, respectively. The training set and validation set R2 of VIs&TFs fine segmentation model were 0.78 and 0.81, RMSE were 4.73 and 4.31, and RPD were 2.26 and 2.05. In summary, the refined semantic segmentation with the inter-ridge background feature effectively improved the accuracy of the wheat scab monitoring model. The monitoring model was realized to classify the diseases, and then to generate the spatial distribution map of wheat scab grade. The disease distribution of wheat scab was obtained to predict the trend of wheat scab diseases. The finding can also provide effective guidance for the variable application in the late stage of wheat cultivation. [ABSTRACT FROM AUTHOR]

With rapid urbanization and the structural transformation of economic development, cultivated land abandonment is becoming more serious, posing great threats and challenges to the security of food production. The rapid and accurate monitoring of cultivated land abandonment is of great value for ensuring regional and national food security. In view of the urgent need for abandoned land monitoring, we proposed a method to identify abandoned land based on the temporal features of optical time series data. First, the NDVI time series of cultivated lands was constructed using multitemporal Sentinel-2 data. The constructed NDVI time series was preprocessed to filter out outliers among the observations. Subsequently, the amplitude feature of the NDVI time series was extracted for cultivated land. We acquired the amplitudes of the abandoned and non-abandoned land samples and generated statistics for the range of amplitudes. We then extracted the optimal threshold of NDVI amplitude for separating abandoned land from non-abandoned land using two steps: initialization and optimization of the threshold. In the initialization step, the maximum NDVI amplitude among the abandoned land samples was set as the initial threshold. In the optimization step, the threshold was decreased in step of 0.01, and the F1-score was calculated iteratively to determine the optimal threshold. With a decrease in the threshold when the F1-score reached its highest value, the corresponding threshold was set as the optimal threshold. Accordingly, the rule for abandoned land recognition was constructed; namely, when the NDVI amplitude was lower than the optimal threshold, cultivated land was classified as abandoned land. Finally, based on the mapping results, landscape pattern indices were calculated to analyze the landscape pattern characteristics of abandoned lands. The proposed method was validated in Potou District, Zhanjiang City, Guangdong Province, China. The main results of the study were as follows: 1) Comparing the NDVI time series curves of abandoned and non-abandoned lands, we found that the NDVI time series curves of abandoned land changed gently, whereas the NDVI time series of non-abandoned land changed with great fluctuation due to the phenological process of crop growth. 2) By iteratively searching for the optimal NDVI amplitude, the optimal segmentation threshold of NDVI amplitude was 0.42 for distinguishing abandoned lands from non-abandoned lands in Potou. We used 100 abandoned and non-abandoned land samples to validate their accuracy. By using this optimal threshold, the overall recognition accuracy of abandoned land reached 91.83%, and the overall extraction accuracy of non-abandoned land reached 90.20%. 3) The calculation results of the landscape pattern index of abandoned lands revealed that these are generally small and have irregular shapes. Abandoned lands were scattered throughout Potou. This study demonstrates that the proposed method can effectively identify abandoned land using the amplitude of NDVI time series. This method can produce highly reliable abandoned land mapping and shows good potential for largescale agricultural applications. [ABSTRACT FROM AUTHOR]

Objectives: The distribution of buildings is an important indicator to measure regional development. Automatic extraction of building information from remote sensing images is of great significance for urban and rural planning. Most existing methods underestimate building details such as boundaries and tiny buildings. Methods: In order to increase attention to building details, a dual-stream detail-concerned network (DSDCNet) is proposed in an encoder-decoder manner. First, a dual-stream feature extraction module is used to extract semantic features and detail-concerned features. They are fed into the decoder consisting of a series of detail refinement modules where detail-concerned features make up for the missing details of semantic features and the semantic features enhance semantic continuity of detail-concerned features. Then, a semantic-detail fusion module is used to fuse and squeeze two refined features. Furthermore, deep supervision is conducted and the multi-level outputs are used in detail-concerned loss function so as to strengthen the supervision of building details. Results: Five mainstream networks are selected for comparison in WHU dataset, ISPRS Vaihingen dataset and a domestic high-resolution dataset. The evaluation results show that DSDCNet has better performance than other networks, especially in F1-score and intersection over union without introducing too much network complexity. Conclusions: DSDCNet not only manages to improve the overall performance of building extraction results, but also effectively maintains the integrity of building boundaries and reduces the missed detection of small buildings. It has better extraction effect on the buildings with small size and complex context. [ABSTRACT FROM AUTHOR]

Objectives: The quantification of debris flow material sources is the basis of scientific evaluation of debris flow scale, degree of disaster and comprehensive management. It is difficult to identify debris flow source under dense vegetation by traditional ground survey and optical remote sensing in lush mountainous areas of southwest China. Methods: Comparatively, airborne light detection and ranging (LiDAR) technology can effectively remove vegetation to obtain high resolution digital elevation model, providing a new solution for debris flow source identification. Taking the Rize basin debris flow in Jiuzhaigou earthquake area as a case study, we propose a new methodology for interpretation debris flow source based on high-resolution LiDAR data. The material sources are divided into slump-mass source, gully sediments source and slope wash source, and we propose geomorphologic identification masks and remote sensing interpretation to regard different types of debris flow source. Results: A total of 155 debris flow material sources in Rize basin debris flow are interpreted, covering an area of 1.06 km2, accounting for 31.56% of the total basin. On this basis, the characteristics and distribution law of various types of material sources are analyzed. Conclusions: The results provides theoretical reference and data support for the quantification of debris flow source and debris flow prevention and risk assessment of the earthquake area in Jiuzhaigou. [ABSTRACT FROM AUTHOR]

Objectives: Remote sensing scene classification provides new possibilities for the application of high-resolution images, and how to effectively realize scene recognition from high-resolution remote sensing images is still an important challenge. The existing scene classification methods only use remote sensing images from one viewpoint for scene recognition, which cannot accurately express the semantic information of complex high-resolution remote sensing images, and the accuracy of scene classification is difficult to be further improved. Methods: To solve this problem, a multi-view scene classification method for remote sensing images is proposed. First, the aerial image and ground image are constructed into a positive and negative image pair, and divided into training dataset, validation dataset and test dataset. Second, a convolutional neural network with fusion multi-scale attention is constructed, and features with fusion attention and strong representation ability are obtained through feature fusion module, so as to integrate different feature information and realize multi-scale feature learning. Third, the trained multi-scale attention network is used to extract features from aerial image and ground image, respectively. Finally, the fused features are used to classify scenes based on the fused features using support vector machine. To demonstrate the performance of the proposed multi-scale attention network, we conduct experiments on two publicly available benchmark datasets − the AiRound and the CV-BrCT datasets. Results: The proposed method achieves remarkable performance, with the highest accuracy of 93.13% in the AiRound dataset and 85.18% in the CV-BrCT dataset, which improves the accuracy of single-view scene classification. Conclusions: The results demonstrate that the complementary information provided by multi‐view images can further improve the performance of remote sensing scene classification. [ABSTRACT FROM AUTHOR]

Objective: Deep convolutional neural network (DCNN) is widely used in automatic road extraction from high-resolution remote sensing images (HRSIs). However, the existing methods make it difficult to model the context relationship of pixels in the predicted results. To solve this problem, some studies have used fully connected conditional random field (FullCRF) to perform secondary optimization of semantic segmentation results combined with context information, but the discontinuity problem of road structure cannot be effectively improved. Methods: In order to improve the integrity of road structure, this paper proposes a short range conditional random filed (SRCRF) model combined with DCNN. SRCRF mainly includes a unary potential function based on road pre-segmentation, a binary potential function based on spectral spatial features, and a K-neighborhood mean field inference algorithm. First, the prior knowledge of road pre-segmentation results is obtained by using the powerful feature extraction capability of DCNN as the unary potential function of SRCRF. Then, the dependence of the binary potential function defined by the linear combination of Gaussian kernel functions on the surrounding nodes is modeled. The binary potential function enables the classification results to have local consistency,that is, adjacent pixels with similar spectral features have the same label. Finally, the K-neighborhood mean field inference algorithm based on the mean-field approximation inference algorithm optimizes the inference range to make full use of the spatial context information and spectral feature context information of the road, and then calculates the optimal label corresponding to each pixel based on the space and spectral feature to optimize the road accurately. The convolution method is adopted to control the inference range of SRCRF within the radius of K in order to improve the proportion of feature vectors. Results: The experimental results show that SRCRF not only alleviates the transition smoothness of FullCRF, but also alleviates the structural discontinuity and incompleteness in the road acquisition results of HRSIs. In Zimbawe-Roads dataset and Cheng-Roads dataset, F1-scores of SRCRF are increased by about 4.01% and 3.73% respectively compared with DCNN, and about 3.25% and 2.28% respectively compared with FullCRF. Conclusions: The proposed SRCRF combines the advantages of DCNN, optimizes the fully connected structure of traditional conditional random fields into the K-neighborhood structure, reduces the inference scope, and improves the proportion of feature vectors. Compared with FullCRF, SRCRF can make better use of image color features and spatial features to accurately optimize the road extraction results of deep learning output. The performance of SRCRF is improved compared with DCNN and FullCRF, and the time is shortened by one order of magnitude compared with FullCRF. In the future, we will further investigate the potential for learning Gaussian features and investigate more complex architecture of conditional random field to better capture global context information. Additionally, we are particularly interested in exploring the application potential of SRCRF in other fields, such as building extraction, vehicle extraction, lake extraction, etc. [ABSTRACT FROM AUTHOR]

Channels often freeze in cold regions in winter. A control system of horizontal freezing always causes a significant decrease in the water delivery capacity of channels. Improper regulation may even lead to disasters, such as ice jams and ice dams. A large number of studies on the channel ice conditions have been conducted to improve the water delivery capacity and safety during the ice period. However, it is still lacking in the scheduling decisions from the verified numerical simulation, particularly for the spatiotemporal density of observation data. Remote sensing has great potential for the channel ice monitoring, due to the wide range of observation and high timeliness. This study aims to explore the suitable method for the remote sensing monitoring of channel ice conditions in cold regions. The open channel was taken from the BeijingShijiazhuang section in the middle route of the south-to-north water transfer project. Three types of feature space datasets were constructed using 11 bands of Sentinel-2 images, including the complete feature, the optimized feature, and the combined feature. These datasets were selected as the inputs for the classification of the support vector machine (SVM), maximum likelihood estimation (MLE), and random forest (RF). Nine classifiers were trained to identify the freezing range of the channel. An experiment was conducted to extract the freezing range of the channel from the images of the Beijuma check-gate. The performance of classification was compared under different classifications and inputs. The results indicated that the nearinfrared (NIR), visible light (R, G, and B), and shortwave infrared (SWIR) were the key bands to recognize the range of the channel freezing. The highest accuracy was achieved in the SVM using the limited sample size, with a producer's accuracy (PA) of 85.10%-87.91% and a commission error (CE) of 10.84%-16.08% under different feature inputs; The accuracy of RF classification was close to that of SVM using complete and optimized feature, with PA=84.67%-86.61%, and CE=13.76%- 14.41%. But the classification was seriously deviated from the reality under combined features; The MLE classification shared the low accuracy under all three types of features, indicating unsuitable for the recognition of channel icing range. Overall, the SVM had a low sensitivity to the feature space, whereas, the high-precision observation was achieved in the range of channel freezing under different feature inputs; The RF shared a high sensitivity to feature space, where the unstable accuracy was observed when the input features changed. Taking the SVM with complete features as an example, the spatiotemporal generalization of the classifier was finally verified for the model. The lowest mapping accuracy of 82.09% and the highest misclassification error of 13.82% were achieved at different times and channel segments. The classification model with the better accuracy can be expected to effectively identify the icing range of the channel. The finding can provide a new approach to monitoring the ice conditions in water transfer projects in cold regions. The application scenarios of satellite remote sensing can also be extended to the strong reference in the field of ice monitoring. [ABSTRACT FROM AUTHOR]

Rice is one of the most important cash crops in the world. Rice classification and mapping are fundamental to assessing food security. The multi-source satellite imagery can be expected to obtain the rice distribution with the advances in remote sensing. A hot research topic is to effectively synthesize data from multiple sensors, in order to improve the accuracy of crop extraction. In particular, the optical and radar data can be combined to classify the rice in the context of cloudy and rainy regions. Previous studies on rice extraction from synthetic aperture radar (SAR) and optical images have achieved good results. However, simple stacking was usually used in feature fusion without considering the interaction between optical and SAR images. Moreover, mostly shallow features, such as texture and geometric features,were extracted and used for classification,rather than high-level semantic features. In this study, a multi-level feature fusion framework was proposed to improve the existing network of deep learning (HRNet). The improved model obtained was named MSATT-HRNet. The dualpolarized SAR imagery of Sentinel-1 and the multispectral optical imagery of Sentinel-2 from the ESA Copernicus project were integrated to extract the rice cultivation areas in Wangcheng District, Changsha City, Hunan Province. The improvements to HRNet mainly included two parts: 1) The convolution group composed of the channel attention mechanism and the maximum pooling was designed to extract the SAR features.The self-attention module was embedded in the basic module of feature extraction in HRNet for extracting features of multispectral optical images. 2) A feature fusion module was designed with the channel and spatial attention, in order to explore the intrinsic complementary relationship between the dual-modal features. Ablation experiments were carried out to verify the improved model. The performance of the MSATT-HRNet network was evaluated to compare the rice extraction with the common deep learning (MCANet, Deeplabv3, and Unet), and the original HRNet with only optical images before improvement. The results showed that a significantly higher overall accuracy was achieved in the rice mapping scheme embedded with the improved feature extraction and fusion module. The multi-source data fusion greatly contributed to balancing various data sources. The overall accuracy and Kappa coefficient of the MSATT-HRNet model were as high as 97.04% and 0.961, respectively. Compared with the MCANet, Deeplabv3, Unet, and original HRNet, the overall accuracy of the model was higher by 6.90, 2.67, 2.98, and 4.85 percentage points, respectively, while the Kappa coefficient was higher by 0.055,0.025,0.030 and 0.041, respectively. The comparison confirmed that the improved model can effectively improve the accuracy of rice extraction. A feasible option was offered to couple deep learning techniques and remote sensing images for the rice mapping in the southern cloudy and rainy region. Nevertheless, some limitations still remain in the application of the improved model due to the availability of optical images of the area. [ABSTRACT FROM AUTHOR]

Wheat, as one of the three major food crops in the world, plays a crucial role in the global food supply system. It is very necessary to accurately and timely predict the winter wheat yield for national food security. Spectral information obtained from remote sensing satellites can be expected to better reflect the growth status of crops, and then predict the yield of largescale crops. Machine learning has also been widely applied in crop classification and yield prediction in recent years, due to its strong capabilities in data mining and analysis. However, the input factors of machine learning are mostly traditional variables, such as soil moisture, meteorology, and land use. It is a high demand to introduce new variables as the input factors for machine learning, in order to improve the accuracy of crop yield estimation. Solar-induced chlorophyll fluorescence (SIF, a byproduct of vegetation photosynthesis) has gradually become a new variable in crop yield prediction, which is sensitive to water and heat stress during crop growth. The data from different time windows can be used to determine the optimal time window and the high accuracy of yield prediction of winter wheat. Meanwhile, the limited data can be used to provide reliable predictions before the winter wheat harvest. In this study, an ideal region was taken to predict the wheat yield per unit area in Henan Province in China. The yield of winter wheat was then predicted using multi-source remote sensing data and ensemble learning. The growth season of winter wheat was divided into 28 time windows. SIF feature variables were introduced to normalize the vegetation index, water index, leaf area index, monthly average temperature, and total monthly precision. Multiple remote sensing indicators and meteorological data were collected from 2003 to 2018, in order to train the yield of winter wheat from 2019 to 2021. Stacking-based ensemble learning was used to fit eight machine learning models, and then combine their predictions. The prediction errors were then reduced to improve the accuracy and generalization of the model. The results show that: (1) SIF features improved the prediction of winter wheat yield; (2) The optimal time window of machine learning was in the period of December to the following May; (3) Compared with the rest, the Stacking-based ensemble learning performed the best, where the coefficient of determination was 0.816, the root mean squared error was 580.36 kg/hm2, and mean absolute error was 476.01 kg/hm2 ; (4) The spatial distribution of the predicted yield was similar to the actual one, with a trend of low in the west and high in the east. This finding can also provide new ideas for predicting crop yield. [ABSTRACT FROM AUTHOR]

: Stripe rust has posed a serious threat to the wheat yield in recent years. It is crucial to breed the wheat varieties resistant to stripe rust. However, the identification of resistance is single and inefficient in traditional breeding. In this study, an efficient identification was proposed to determine the different resistance grades to the stripe rust using the time series of vegetation index response to wheat canopy. An unmanned aerial vehicle (UAV) was utilized to collect multi-temporal spectral images of the canopy in the naturally occurring breeding populations of colony wheat (600 samples in total, 516 genotypes). Six sensitive features were screened for the severity of stripe rust disease using Random Forest and ReliefF algorithms: normalized pigment chlorophyll index (NPCI), woebbecke index (WI), chlorophyll index rededge (CIrededge), (green atmospherically resistant index GARI), normalized difference vi (NDVI), and chlorophyll index green (CIgreen). These indices were verified as sensitive features. The severity of stripe rust disease incidence was dynamically characterized using the time series of these indices in the test population. The support vector machine (SVM) was used to establish a classification model for the severity grade of stripe rust disease, according to the sensitive features. There was a very small difference in the performance of the test set and the unscreened original features, indicating the effectiveness of the screened sensitive features. The time series of six sensitive traits was observed in the samples of different resistance grades. It was found that there were no significant differences in the CIgreen and CIrededge among the samples with the different resistance grades. This indicated that the saamples were not applicable to categorize the resistance grades to stripe rust. The differences exhibited by GARI, NDVI, NPCI and WI were used to classify the resistance grades to stripe rust. General machine learning cannot capture the smaller differences of feature variation in the samples with the different resistance grades. Therefore, an improved mode was proposed to extract the features from two-dimensional images that transformed vegetation index time series, in order to realize the classification of stripe rust resistance grade. Four time-series vegetation indices (NPCI, GARI, NDVI, and WI) were better distinguished the different disease resistance grades among the sensitive features, and then used to generate the Gramian Angular Summation Field (GASF) images by the Gramian Angular Field. Data augmentation was performed on the dataset to equalize the number of samples in each resistance grade. Each dataset had a total of 1 040 samples, and was then divided into four grades of stripe rust resistance, where each grade contained 260 sample images, while each dataset was divided into training, validation, and testing sets in the ratio of 6:2:2. DenseNet121 model was separately trained using each dataset, in order to classify the various stripe rust resistance. A better performance was achieved in the classification models with the GASF_NPCI and GASF_WI as the input features, compared with the GASF_GARI and GASF_NDVI. The model with the GASF_NPCI as a feature was slightly less effective in distinguishing the samples with the resistance grades R and MR, where the precision and recall were relatively low. There was no difference in the models with the GASF_WI for the precision and recall of the samples that predicted each stripe rust resistance grade. In the F1 scores of the test set, the different vegetation indices on the resistance grades of stripe rust in colony wheat were ranked in the order of NPCI, WI, GARI, NDVI. The classification model with the GASF_NPCI was the most effective in the test set, with an F1 score of up to 0.833. There was a better distinction of differences in the stripe rust resistance grades among different varieties (lines) of population wheat. The grades of wheat stripe rust resistance were fully identified using time series of spectral vegetation index. Meanwhile, the finding can also provide a strong reference for the disease resistance breeding of crops. [ABSTRACT FROM AUTHOR]

Since the objects in remote sensing images have the characteristics of arbitrary direction, dense distribution and large scale differences, object detection in remote sensing images has become a challenging problem. Aiming at this problem, this paper systematically reviewed the work related to rotating object detection in deep learning remote sensing images in recent three years. Firstly, this paper introduced the representation method and characteristics of rotated bounding box. Then, it analyzed the existing methods of rotating object detection in remote sensing images from four aspects: the feature extraction network, the generation of rotating anchor and candidate bounding box, the label allocation and sampling strategy, and the loss function. Next, it introduced the commonly used remote sensing image datasets for rotated object detection, and compared the performance of different algorithms. Finally, this paper prospected rotating object detection in remote sensing images. [ABSTRACT FROM AUTHOR]

Low spatial resolution is one of the main bottlenecks restricting the fine application of remote sensing parameters. Spatial downscaling is an effective way to improve their spatial resolution and application capabilities. Researchers have developed various downscaling methods for different remote sensing parameters. However, a unified and general method classification system has not yet been formed. Based on the in-depth commonality analysis of the downscaling of various parameters, this paper systematically summarizes the spatial downscaling methods from the perspective of data fusion, taking into account the required complementary information. Four types of downscaling methods are summarized: Multi-parameter fusion, spatiotemporal fusion, remote sensing-model fusion, and super-resolution reconstruction implicit fusion. The advantages, disadvantages, and applicable scenarios of various methods are analyzed, and the development trend of spatial downscaling method research is discussed. It can provide theoretical and technical support for improving the refined application ability of remote sensing data. [ABSTRACT FROM AUTHOR]

Large-scale non-grain cultivation has posed a serious risk to national grain security. Non-grain cultivated land can also deteriorate the ecological environment, such as soil quality, greenhouse gas emissions and agricultural pollution. Therefore, it is of great significance for the accurate monitoring of non-grain cultivated land using remote sensing data. However, the existing research has focused mainly on the distribution of specific crop types or cash crop expansion. It is still lacking in the non-grain croplands and their spatial distribution. In this study, the remote sensing time series were selected to identify the non-grain croplands in Gaocheng District, Shijiazhuang City, Hebei Province, China. The spatial distribution of non-grain cropland was also obtained from 2019 to 2022. The results demonstrated that: 1) Both pixel and object-oriented machine learning showed a high classification accuracy to identify non-grain cultivated land. Spectral bands were integrated to extract the phenological features from the NDVI time series. The growth status and temporal patterns of different crops were captured to improve the classification accuracy. The low accuracy was also found in the time series matching on the open-field vegetables, greenhouses, and uncultivated croplands, due to the high within-class variability. A better performance was achieved to distinguish between grain crops and other land cover classes. 2) The pixel approach was more sensitive to the land cover at the pixel level, thus capturing the differences among various types of land cover. The object approach effectively reduced the salt-and-pepper noise, and then significantly improved the confusion among land cover categories with the high within-class variability. 3) The object machine learning exhibited the highest accuracy for the classification and identification of non-grain cropland conversion, with overall accuracies of 87.00% and 81.00% for two growing seasons, respectively. 4) The annual non-grain cropland area was 2 753.09 hectares, with the orchards accounting for up to 62.55% of the total. The seasonal non-grain analysis showed that the autumn non-grain area (3 174.86 hectares) was significantly higher than the summer ones (1 060.27 hectares). In conclusion, the different remote sensing time series can be expected to accurately identify and monitor the non-grain cultivated land. The specific recommendations were selected for the practical applications, according to the application needs and data availability. Machine learning can be selected to fuse the harmonic and phenological features, when the specific types of non-grain cultivated land need to be distinguished. However, the time series matching can also be used for the limited sample sizes without considering the specific non-grain cultivated land. [ABSTRACT FROM AUTHOR]

The planting structure and spatial distribution of crops are the basis for the estimation of crop losses in flood disasters. In order to meet the accuracy requirements for rapid assessment of crop losses in flood disasters, it is urgent to build a real-time crop survey and disaster assessment method that combines remote sensing technology with ground surveys. Taking the rainstorm and flood disaster in Henan Province in 2021 as an example, this paper comprehensively used satellite remote sensing data and real ground object sample data obtained from field research to carry out a research on the fine classification of crops in Weihui City, Xinxiang, Henan Province, combined with the flood inundation duration calculation results based on remote sensing data, to assess the damage of crops in this flood disaster event in Weihui City. The results show that: a) The main types of flood disasters in Weihui City are high water level intrusion floods in rivers and waterlogging and stagnation in low-lying areas. b) The types of disaster-affected features are corn, the main crops, green forests around roads, and urban areas of Weihui City. The disaster assessment method for flood crops based on multi-source remote sensing data and real ground object samples proposed in this paper has high accuracy and timeliness. [ABSTRACT FROM AUTHOR]

Cultivated land makes significant contribution to the land use of arid areas. Analyzing the spatiotemporal evolution of its production – ecological effects can reveal the role of cultivated land changes in regional food production and ecological environmental protection, and provide a reference for achieving synergistic production –ecological protection. Based on land use classification data from four periods during 1990, 2000, 2010, and 2020，we used the land use transfer matrix，production hypothesis，and remote sensing ecological index（RSEI）models to reveal production–ecological effects under spatiotemporal evolution of cultivated land in Hotan Region. From 1990 to 2020，the cultivated land area increased by 1 702.63 km2 ; eight counties showed the same trends with different absolute values，in cultivated land area. Reclamation of unused land，forests，and grassland were the main sources of newly cultivated land; the return of cultivated land to forests and grassland，expansion of construction land，and abandonment of cultivated land were the main reasons for decreases in the cultivated land. The level of cultivated land production increased. The proportion of unproductive cultivated land decreased by 12 percentage points and the proportion of medium and medium–high productive cultivated land increased by 14 percentage points. In terms of production effect of cultivated land，the positive effect of unchanged area production（55%）was greater than the negative effect（2%），and the transfer effect of cultivated land（43%）effectively compensated for the lost production effect area. However，the fixed negative effect area gradually increased，whereas the fixed positive effect area gradually decreased. The transfer effect area showed a fluctuating downward trend in terms of the effect of each stage. The study area is mainly dominated by level 3 ecological quality cultivated land. The ecological and environmental quality of cultivated land deteriorated over the study period，with the area of fixed cultivated land decreasing by more than twice the area of improvement，and the area of transferred cultivated land decreasing by nearly 1.4 times the area of improvement. In the phase change，the environmental quality of cultivated land improved in the first 20 years and deteriorated in the last 10 years. This deterioration was caused by the linear deterioration of the environmental quality of fixed and transferred cultivated land. The results of this study demonstrate that changes in cultivated land in the study area have contributed to an increase in regional production， which has in turn led to a gradual deterioration in the ecological quality of cultivated land. Therefore，future spatial planning and land management should emphasize the development of production–ecological integration to optimize the allocation of land resources. [ABSTRACT FROM AUTHOR]

In the past few decades, earth observation has experienced rapid development, which provides necessary theoretical knowledge, research means and methods for global sustainable development. By searching the Web of Science core collection database and using CiteSpace, a knowledge map tool, this study conducted a bibliometric analysis of sustainable development research based on earth observation from 1991 to 2021. This study analyzed the research progress based on the literature output trends, publishing regions and cooperation network, high-frequency terms and terms with the strongest citation bursts, and proposed future research directions. The results show that: ①The number of articles published in the field of "earth observation technology and sustainable development" is generally on the rise, which can be divided into germination stage, growth stage and rapid development stage. ②There is close international cooperation in the study of earth observation technology and sustainable development, among which, United States and Germany are in a leading position in terms of the number and influence of published paper. These countries are rich in scientific research. China and other developing countries have a dominant position in the number of publications, but their international influence still needs to be improved. ③At present, research focuses on the environmental protection related contents such as land use/land cover change, climate change, natural resources and water resources. In recent years, the research field has been constantly integrating artificial intelligence, mathematical statistics and other emerging methods, and the research content tends to be diversified. In the future research, it is necessary to fully achieve the supporting role of earth observation technology and contribute to the achievement of the Sustainable Development Goals through a Decade of Action, and to deepen the study of earth observation data and facilitate data sharing in order to open up a new path for wide application. The combination of earth observation data with traditional statistical data should be strengthened to achieve the fully functioning of the potential of earth observation in SDGs monitoring. We can focus on international cooperation and take global action. These can help us to provide effective strategic and decision reference for the realization of the Sustainable Development Goals. [ABSTRACT FROM AUTHOR]

Most remote sensing image data is inevitably contaminated by clouds, resulting in data failure. Therefore, detecting clouds is a necessary preprocessing step. With the rapid development of space technology, more portable satellites have been designed to equip these micro satellites with remote sensing image preprocessing models. Designing a high-precision and low computational power requirement lightweight cloud and cloud shadow detection network model is of great significance. In response to the above issues, a lightweight M-shaped network (L-MNet) based on deep separable convolution was proposed. The L-MNet network model was introduced depthwise separable convolution (DS-Conv) on the basis of the M-shaped network (M-Net) model. A depthwise separable convolutional block (DS Conv Block) was designed to reduce the complexity and computational complexity of the algorithm. The results show that the proposed method can effectively reduce the model size and computational complexity of pixel level cloud detection while ensuring detection accuracy, and is helpful in achieving the task of small satellite in orbit cloud detection. [ABSTRACT FROM AUTHOR]

Spatiotemporal resolution of remote sensing precipitation products has been continuously improved in recent years. New possibilities can be provided to estimate the spatially continuous rainfall erosivity at the regional scale. However, previous studies have often failed to consider the differences between these products and observation data, as well as the potential for bias correction. In this study, the remote sensing precipitation products were used to estimate the rainfall erosivity in Guangdong Province, China. The precipitation data was collected hourly from 86 meteorological stations from 2001 to 2020. Two IMERG remote sensing precipitation products were selected, including the GPM IMERG Final Precipitation L3 Half Hourly 0.1 degree × 0.1 degree V06 (GPM_3IMERGHH, 0.1°, every 30-min) and GPM IMERG Final Precipitation L3 1 day 0.1 degree × 0.1 degree V06 (GPM_3IMERGDF, 0.1°, daily). The accuracy of two IMERG remote sensing precipitation products in estimating rainfall erosivity at semi-monthly, annual, multi-year average, and extreme event scales was then evaluated to quantify the bias. Fitting and bias correction were utilized to determine the optimal rainfall erosivity using remote sensing precipitation data. The EI30 was an indicative measure of event rainfall erosivity derived from hourly rainfall data. Fitting revisions of the EI30 were estimated from 5-minute rainfall data collected from three meteorological stations. The results showed that neither product was suitable for the direct estimation of rainfall erosivity with low accuracy at different time scales. Pearson correlation coefficients (CC) and Kling-Gupta efficiency (KGE) were less than 0.63 and 0.51 on a semi-monthly scale, indicating low estimation accuracy and small correlation. The KGE values were less than 0.5 on an annual scale, whereas, the CC values were higher than those on a semi-monthly scale, all exceeding 0.63. The KGE values were still less than 0.5 on the multi-year average and extreme event rainfall erosivity. But the CC was greater than or equal to 0.78. There was a strong correlation between the estimated and the observed values, indicating the potential for bias correction. Therefore, the linear models were used to correct the IMERG estimates. Among them, the correction models were y=2.0017x-1373 and y=1.6916x-4, respectively, for the multi-year average annual rainfall erosivity (R-factor) and 10-year storm EI30 (EI30 that occurs once in 10 years). Cross-validation results showed that the corrected GPM_3IMERGHH estimates of R-factor and 10-year storm EI30 had the KGE values of 0.79 and 0.64, respectively, which were superior to the spatial interpolation using station daily rainfall data (KGE values of 0.60 and 0.59, respectively), and similar to the station hourly rainfall data (KGE values of 0.77 and 0.66, respectively). R-factor and 10-year storm EI30 were necessary to require for the areas of Guangdong Province without observation stations, in order to map the spatial distribution of these factors. Specifically, IMERG remote sensing precipitation products were used to estimate the rainfall erosivity with the bias corrections. The accuracy of the corrected estimates was superior to that using spatial interpolation with the station daily rainfall data, and similar to that with the station hourly one. The findings can effectively demonstrate the substantial potential of the remote sensing precipitation products in the field of soil water erosion, indicating the broad range of future possibilities [ABSTRACT FROM AUTHOR]

In recent years, the application scenarios of satellite remote sensing images have become increasingly diverse. However, due to limited collection equipment and cost constraints, the images obtained by satellite sensors usually do not have sufficient resolution and are not uniformly distributed, which is difficult to distinguish some clustered objects. Low-resolution remote sensing images are not suitable for semantic understanding tasks such as object detection and recognition to accurately locate and classify all objects. Compared to obtaining complete high-resolution remote sensing images at once, existing remote sensing images usually have relatively clear high resolution in local areas and sufficient detailed information for distinguishing object groups. Traditional remote sensing image super-resolution reconstruction methods mainly focus on the global features of remote sensing images, expanding resolution and pixels based on global features of images, while ignoring the details of remote sensing images. To address this problem, this paper proposes a method that introduces detailed information about local clustered object areas in local images in the reconstruction of complete remote sensing images. Specifically, the proposed method uses a multi-level neural network to extract image features of different scales and then uses residual learning to merge and reconstruct these features. In the experiments of this paper, the proposed method achieved better visual effects and numerical results compared to several existing remote sensing image super-resolution reconstruction methods. This indicates that the proposed method can effectively utilize the pixel information of local images, significantly improve the details of global remote sensing images and optimize the recognition capability of the group objects area, and enhance the quality and availability of remote sensing images in a low-cost way. [ABSTRACT FROM AUTHOR]

Gully erosion is one of the most important indicators of soil loss in the dry-hot valley of the Jinsha River. Serious land degradation has threatened regional agriculture and food security. It is of great significance for the quantitative evaluation of gully erosion and land protection. In this study, the erosion rate was determined from the morphology evolution and development of the gully. The research area was selected as the gully geomorphic watershed near Jinlei National Soil and Water Conservation Science and Technology Demonstration Zone in Yuanmou Dry-Hot Valley. 35 gullies were selected to extract the gully shape parameters from the remote sensing images of unmanned aerial vehicles (UAV). Morphological characteristics were analyzed to calculate the gully erosion rate combined with high-resolution remote sensing images from 2015 to 2022. The results show that: 1) The straight length, curve length, top width, bottom width, depth, area and volume of gully in the study area were 25.21-180.43 m, 25.73-203.01 m, 10.76-51.76 m, 3.56-26.85 m, 1.22-11.92 m, 257.50-8987.88 m² and 142.64-19479.25 m³, respectively. There was a great variation in the morphological parameters of the gully. The length of the gully was less restricted by slope than other areas during development. 2) The ratio of length to top width of gully (L/TW) ranged from 1.35 to 6.38, with an average value of 2.78. There was no significant gully narrowness. The strong lateral erosion occurred at the same time as retrogressive erosion. The ratio of the bottom width to the top width of the gully (BW/TW) ranged from 0.24 to 0.59. The flushing channels were mainly in the "V" and "V-U" shape. The ratio of top width to depth of gully (TW/D) ranged from 2.56 to 26.29, with an average value of 6.04. The lateral erosion rate was much higher than that of the downward. There was a great difference in the cross section of the gully from gully head to mouth. The tortuosity index of gully (TI) ranged from 1 to 1.13, with an average value of 1.05. There was a small bending degree of the gully. 3) 54.29% of the 35 gullies experienced an increase in length, where the retrogressive erosion rate of the gully was 0.02-1.10 m/a (average 0.34 m/a). 77.14% of the gullies experienced an increase in width, where the lateral erosion rate ranged from 0.07 to 1.10m/a (average 0.36m/a). 80.56% of the gullies experienced an increase in area, and the growth rate of the gully area was 1.64 - 105.40m² /a (average 23.44m² /a). The optimal fitting was obtained for the power function of 35 gullies area and volume: V=0.058 4A 1.405, with R² =0.946 1. The volume of the gully was calculated to combine the gully area from the GF-2 remote sensing image. The volume erosion rate of the gully ranged from 1.26 to 339.42 m³ /a (average 51.21 m³ /a). Both the number and rate of lateral erosion gullies were greater than those of retrogressive erosion gullies. Lateral erosion greatly contributed to the growth of the gully area and volume. The findings can provide a strong reference for the development and evolution of large-scale gullies, in order to quantitatively evaluate the erosion in the Dry-Hot Valley of Jinsha River. [ABSTRACT FROM AUTHOR]

The timely and accurate extraction of greenhouse instance (GI) is of significant practical importance to estimate the vegetable cultivation areas and yield prediction. Deep learnings (driven by knowledge learned from large-scale sample data) can be expected to adaptively explore the features of image data, compared with the traditional image analysis, such as unsupervised, supervised, and object-oriented classification. The end-to-end accurate extraction of GI information can also ensure the model generalization for the less manual intervention. However, there are still two challenges to identify the GI using deep learning. One is that the multiple GI can be mistakenly assumed as a continuous distribution in the dense areas of GI, leading to the segmentation errors. Another is that the degradation of performance could occur, when transferring the GI model to the large-scale spatial context. In this study, the region-boundary instance extraction was proposed using convolutional neural networks (CNNs) and morphological post-processing. At first, a region-boundary multi-class model was constructed to generate the boundary auxiliary labels of greenhouse. The network loss function was modified to enhance the boundary information recognition, and then to facilitate the removal of greenhouse boundaries from the recognition, thereby achieving the extraction of GI. Subsequently, the high-precision GI data was obtained using morphological operations, such as the instance object dilation and the minimum bounding rectangles. The high-resolution three-band remote sensing data was collected from Shouguang, Shandong Province, in order to train the base model. Three transfer modes (namely, pure transfer, scale-adaptive, and model fine-tuning mode) were then explored in five transfer research areas, including Xinjiang, Liaoning, Yunnan, Hubei, and Zhejiang Province. The accuracy of GI extraction was evaluated using common semantic segmentation performance metrics, unit intersection over union (UIoU), and instance recall rate (IRR). The research results indicate that the UNet was better suited to construct as the "Region-Boundary" multi-class model, compared with the semantic segmentation networks, such as PSPNet, DeeplabV3+, and HRNet. The higher semantic accuracy was achieved in the UNet with the UIoU and IRR of 2.43 and 2.91 percentage points higher, respectively, compared with the overall suboptimal HRNet. Furthermore, two morphological post-processing operations (instance dilation and the minimum bounding rectangle recognition) were simultaneously introduced to increase the UIoU and IRR by 10.53 and 1.44 percentage points, respectively. The scale adaptation mode was then adopted to adjust the input image resolution. The UIoU and IRR were improved from 3.02 to 22.71 percentage points, and from 2.40 to 30.33 percentage points, respectively, in all test datasets of migration areas, compared with the simple migration mode. The fine-tuning of the base model was utilized to adjust the input image resolution in the model fine-tuning mode. The UIoU and IRR were improved ranging from 37.76 to 50.67 percentage points, and from 46.04 to 76.87 percentage points. The higher accuracy was achieved in the GI recognition, with the UIoU and IRR of 13.64 and 14.18 percentage points higher than the conventional approaches, respectively. Simultaneously, the model transfer was applied to select the different migration modes, according to the scene differences between the predicted and training regions. The automated mapping of GI can be expected to efficiently and accurately extract the GI information over the large-scale areas. The finding can provide the information support to the intelligent construction of agricultural facilities. [ABSTRACT FROM AUTHOR]

Thermal infrared imaging can be expected to rapidly and cost-saving monitor the soil moisture content in the largescale farmland. But it is still unclear on the impact of lighting conditions on thermal infrared images. This study aims to explore the impact of direct sunlight or shadow occlusion of the crop canopy and soil on the UAV thermal infrared remote sensing, in order to diagnose the crop water stress and monitor soil moisture content. Summer corn with different irrigation treatments was taken as the research object. The thermal infrared images were divided into four parts: illuminated canopy, shaded canopy, illuminated soil, and shaded soil. The light temperature was extracted from the higher temperature, whereas, the shadow temperature was assumed as the lower temperature. Temperature extraction was used to calculate the 11:00, 13:00, and 15:00 canopy temperature difference (difference between canopy temperature and atmospheric temperature, ΔT), crop water stress index (CWSI), and evaporative fraction (ratio of latent heat flux to effective energy, EF). A comparison was made on three changes in the monitoring effect of the index on soil moisture content after using light temperature (ΔTL, CWSIL, EFL) and shadow temperature (ΔTS, CWSIS, EFS ) at different times. The results show that: 1) There was the variation in the monitoring effects of the three indices over time. The EF monitoring effect was better at 11:00 and 15:00, the crop water stress index monitoring effect was better at 13:00. There was the less change in the ΔT monitoring over time. The temperature index should be selected for monitoring soil moisture content using field conditions and the time of flying the drone; 2) The monitoring effect was improved the most at 11:00 in the jointing period, after distinguishing light temperature and shadow temperature. The R² of EF, EFS, and EFL were 0.54, 0.65, and 0.78, respectively. The R² of CWSI, CWSIS, and CWSIL were 0.47, 0.64, and 0.70, respectively. There was no significant light temperature in the period of tasseling and filling, whereas, the index monitoring effect was significantly reduced using shadow temperature. There was the largest decrease in the CWSIS at 13:00, compared with the CWSI, where the R² decreases were 0.11 and 0.06, respectively. Therefore, it was very necessary to choose the clear and cloudless weather and avoid cloudy days, when shooting thermal infrared images. Furthermore, the impact of lighting conditions on thermal infrared images was also change with the growth period; 3) The best monitoring soil moisture content was obtained using EFL at 11:00 in the jointing and tasseling period, and CWSI at 13:00 in the filling period, where the R² of predicting soil moisture content were 0.75, 0.75, and 0.89, respectively. In addition, the soil moisture content was also dominated the monitoring effect. When the soil volumetric moisture content was around 0.23, both EFL and CWSI shared the more accurate prediction. Different time points, light temperature and shade temperature were utilized to analyze the monitoring effect of three temperature indexes on soil moisture content. The better monitoring time and index were determined in the three growth periods of summer corn. The finding can provide a strong reference for the UAV thermal infrared monitoring of soil moisture content. [ABSTRACT FROM AUTHOR]

Objective: Point cloud density is an important parameter of light detection and ranging(LiDAR) technology. Point cloud density has an important impact on the extraction of remote sensing retrieval index for forest. Methods: The experimental data, sized by 1 600 m×1 450 m, had been obtained by unmanned aerial vehicle (UAV) LiDAR and thinned by the graded random thinning method in order to simulate different point cloud density during actual operation, which was used to extract the remote sensing retrieval index for forest such as canopy closure, gap fraction, leaf area index, height quantile variables and density quantile variables. Then these parameters were used to make difference comparison with the indexes extracted through raw data. Results: (1) The lower the point cloud density is, the lower the extracted canopy closure is slightly, while the extracted gap fraction is slightly increased. The point cloud density has little influence on the extracted canopy closure and gap fraction. (2) When the point cloud density is high, it has little impact on leaf area index, but when the point cloud density is small, it has a great impact on leaf area index, and some areas may have sudden changes on leaf area index. (3) When the point cloud density is large, the effect of point cloud density on height and density quantile variables is not obvious, but when the point cloud density drops to 3.6 point/m², there may be sudden changes in density and height density quantile variables in some areas. Conclusions: In short, the point cloud density has an important impact on the description of forest structural characteristics. The appropriate point cloud density is conducive to describe the forest structure morphology more accurately, but the low point cloud density affects the extraction of remote sensing retrieval index for forest. This study has certain guidance and reference for selection of point cloud density to estimate the remote sensing retrieval index with UAV LiDAR on forestry. [ABSTRACT FROM AUTHOR]

Geological disasters affect every continent with a large population, especially in mountainous countries or regions, and its disasters are obviously high, hidden and sudden. Knowing the history and status of disasters in advance through geological surveys is of great significance for the final realization of potential disaster identification and early warning. At present, it is difficult for traditional artificial ground survey methods to discover and identify major geological hazards and hidden dangers in complex mountainous areas such as dense vegetation coverage or high and steep terrain. As a multi-functional comprehensive detection technology, aerial remote sensing can efficiently obtain the development and distribution characteristics of geological hazards and the temporal and spatial evolution laws due to its unique field of view and not being restricted by ground conditions. First, we summarize the types and development trends of aerial remote sensing platforms commonly used in the field of geological disasters, and analyze the advantages of different load sensor information processing technologies and the main problems of geological disasters. Second, the key research achievements of aerial remote sensing technology in five application stages of geological disaster basic terrain mapping, early identification, investigation and evaluation, medium and long-term monitoring, and emergency response are reviewed, and the requirements and advantages and disadvantages of various technical methods at different stages are discussed. The shortcomings of the application of aerial remote sensing technology in the field of geological hazards are summarized, and the future development trends and suggestions are clarified. [ABSTRACT FROM AUTHOR]

A great challenge has been found to limit the per capita cultivated land and the scarcity of land resources in China. However, the cultivated land can be abandoned across the country with the rapid growth of industrialization and urbanization. It is crucial to accurately and rapidly monitor the abandoned farmland with high accuracy for farmland protection. Land use can also be evaluated to guarantee national food security for the decision-making on modern agriculture. Among them, there is fragmented farmland with a wide distribution in the Loess Plateau of China. The complex terrain has made some challenge to investigate and monitor the abandoned farmland. This study aims to monitor the abandoned farmland on the loess plateau using feature optimization. Data sources were collected from the Chinese satellite images with high spatial resolution and multispectrum. Some feature factors were selected to effectively distinguish the cultivated land from other land types. The prior knowledge was combined with the distinct separability, including the spectral, phenological, topographic and texture features. The spectral feature differences were statistically analyzed from May to October. Spectral time series feature factors were developed to prove the effectiveness and superiority of the schemes. Five types of experimental scheme were also constructed for comparison. The binary particle swarm optimization (BPSO) coupled with the genetic algorithm (GA) was selected to modify the index factor for the optimal selection of feature factors in classification. Random forest (RF) classification and detection were performed post-classification. The spatiotemporal distribution was obtained for the abandoned farmland at field scale between 2011 and 2022. The results showed that the feature optimization improved the accuracy and efficiency in the classification of remotely sensed images with the high spatial resolution. According to an importance assessment for all features, NDVI was found to have the highest MDA score. The final preferred feature was ranked in the order of the green band, red band, near-infrared band, (b1-b2)/(b1+b2), Slope, NDVI, Variance and Contrast. The combination of GA and BPSO exhibited high efficiency in the global search and search speed. The Optimum Index Factor effectively evaluated the overall combination. Feature optimization can be expected to avoid information redundancy and over-fitting issues. The field validation was performed using image interpretation and field investigation. It infers that the abandoned farmland at field scale was recognized from the high spatial resolution images of Chinese satellites with an accuracy as high as 92.48%. There was a 96% overlapping ratio between patches from field investigation and those extracted from images. Among them, the recognition rate of fragmented patches with areas as small as 200-400 m² reached 90.11%. The classification showed that there was abandoned farmland with a total area of 5 492.51 hm² between 2011 and 2022, indicating the maximum abandoned percentage (21.09%) in 2011-2014 and the minimum abandoned percentage (0.99%) in 2020-2022. The large-scale pattern of abandonment of farmland was gradually shifted into a smaller fragmented one by peasant households over this period. As such, the abandoned farmland was monitored at a field scale on the loess plateau region using remote sensing. The finding can provide a strong reference to explore the abandoned farmland using high spatial resolution images of Chinese satellites. [ABSTRACT FROM AUTHOR]

Solar greenhouses have been widely used in modern precision agriculture in China. However, potential negative environmental impacts can also be introduced, including atmospheric pollution, soil contamination, and water pollution. Great challenges have been posed on the rational management of farmlands and resource utilization. Meanwhile, the manual collection of visual data cannot fully meet the rapid acquisition of information from large-scale greenhouse facilities. In this study, Chinese solar greenhouses (CSGs) were focused on detecting the damage levels from the remote sensing images using an improved YOLO-DLSG (YOLO-damage level of solar greenhouse) object detection model. Comparative experiments were conducted using high-resolution solar greenhouse remote sensing image datasets. A dataset was collected with the 3235 images of solar greenhouses and then subjected to data augmentation. The classification standards were also combined with the damage indicators specific to solar greenhouses. The damage levels were then established from the high-resolution remote sensing satellite images. The YOLOv5s lightweight improvement was employed to reconstruct the YOLOv5s backbone with the lightweight MobileNetV3 architecture. This restructure reduced the channel dimensions and network parameter volume. Furthermore, the upsampling operator in the neck layers was replaced and then updated. A lightweight upsampling operator was selected as CARAFE (content-aware reassembly feature embedding), using an upsampled kernel prediction module. The feature recombination module was reconstructed in the upsample structure. EVCBlock (explicit visual center block) was utilized to further reduce the computational complexity for the high accuracy of detection on rotating objects. Finally, the CIoU (complete intersection over union) loss function for the target bounding box regression was replaced with the EIoU (Enhanced Intersection over Union) loss, in order to accelerate the convergence of the prediction box regression loss function for the high accuracy of prediction box regression. The experimental results show that the improved model reduced the number of parameters and floating-point operations per second by 17.91 and 15.19 percentage points, with an increase of 0.4 and 0.8 percentage points in the precision and mean average precision, respectively. Compared with the current mainstream lightweight models, such as SSD, Faster R-CNN, Centernet, YOLOv3-spp, and YOLOv3-tiny, the improved YOLO-DLSG model demonstrated superior performance, in terms of accuracy, recall, detection speed, and model lightweight. Data collection and accuracy validation were conducted in the Yangling Agricultural Demonstration Zone in Shaanxi Province of China. An average recognition accuracy of 84.00% was achieved in the improved model. A challenging task was to determine the distribution, quantity, and extent of greenhouse damage in a cost-effective, efficient, and user-friendly manner. This finding can provide a simple and effective way for the rapid location and automatic quantity extraction of solar greenhouses with different damage levels. [ABSTRACT FROM AUTHOR]

In order to quickly and efficiently monitor the outbreaks of Dendrolimus superans, the 8th forest compartment of South Management Office Forest Farm in Heilongjiang Province was taken as the research area, while the Sentinel-2 remote sensing image in 2018 was used as data source to identify the Dendrolimus superans infestation area in the forest compartment. The original spectral features (8), spectral index features (12) and the texture features (8) were extracted from the preprocessed image. Based on ANOVA and XGBoost classifiers, all features were dimensionally reduced and sorted by importance. The ensemble learning classification algorithm (Random Forest classifier and XGBoost classifier) was used to identify pest areas and compare their accuracy. The results showed that: (1) the XGBoost model with the top 14 important features was the most ideal for the identification of pest areas, and the overall accuracy reached to 95% (Kappa coefficient = 86%), which were higher than the 93% of Random Forest (the top 10 features in order of importance); (2) the top 14 feature names were: EVI1, Mean, MTCI, GNDVI, Variance, B4, B2, Homogeneity, B3, CRI1, EVI2, B8, B5 and CRE. This method can achieve efficient identification of Dendrolimus superans infestation areas, which can provide a basis for decision-making on pest control in northeast forest. [ABSTRACT FROM AUTHOR]

In order to solve lack of comprehensive assessment of forest ecosystem service functions on ecological management needs in Mudanjiang City, taking the forest ecosystem of Mudanjiang City in Heilongjiang Province as the research object, referring to the 'Technical Guidelines for Delineation of Ecological Protection Red Line' and the local physical and geographical characteristics, a model for evaluating the importance of forest ecosystem service functions was constructed through four ecological service function evaluation indicators: water conservation, soil and water conservation, wind prevention and sand fixation and biodiversity conservation. The results showed that the forest coverage rate in Mudanjiang City reached 67. 40% and the average net primary productivity of forest vegetation was (730. 06 ± 56. 64) g / (m 2·a) from 2000 to 2020, which showing a significant increasing trend. From the single evaluation index, the importance of soil and water conservation function for the local forest ecosystem was the highest. From the comprehensive evaluation results, the area of extremely important grade of the ecosystem service function was the largest, accounting for 48. 74% of the total forest area, which was concentrated in the southwest of Ning'an City and the south of Muling City. The results could provide scientific basis for optimizing local ecological protection strategies and coordinating regional development. [ABSTRACT FROM AUTHOR]

Soil texture affects soil water content and fertilizer and air permeability, and drives multiple physical and chemical processes occurring in the soil. Developing efficient and fast remote sensing technologies to predict the spatial distribution of soil texture has important theoretical and practical applications in soil quality evaluation and agricultural production planning. This paper discusses the application of remote sensing data, methods and models to predict soil texture, and the use of auxiliary data such as radar, terrain, and vegetation index data for soil texture prediction by remote sensing. The paper also proposes three methods for predicting the spatial distribution of soil texture based on remote sensing characteristics; namely, spectral response, characteristic wavelength selection, and remote sensing interpretation. It also focuses on the application effects of statistics, geostatistics, and machine learning models combined with remote sensing on spatial prediction of soil texture. The advantages, disadvantages, and application of several typical methods are compared, and the application conditions and accuracy verification methods of remote sensing prediction of soil texture are analyzed. Future research needs to focus on in-depth extraction of various remote sensing spectral features, acquisition of multiple types of environmental variables by remote sensing technologies, and development of a multi-algorithm hybrid model combining soil physical properties and data-driven machine learning features, so as to provide a basis and technical support for the spatial prediction of soil texture at different regional scales [ABSTRACT FROM AUTHOR]

Ephemeral and permanent gullies are vital components of erosion gully system, and the existence of a significant correlation between them has become one of the critical scientific questions in the control of gully erosion on the Loess Plateau in the new era. This study aims to explore the effect of ephemeral gullies above the head of permanent gullies on the occurrence and development of the connected permanent gullies in the Loess Plateau. Three typical small watersheds were selected as the research area located in the upper, middle, and lower reaches of the Chabagou watershed in Zizhou, Shaanxi Province. This study systematically analyzes the current connectivity between ephemeral and permanent gullies in the hilly-gully region of Loess Plateau and the impact of this connectivity on the formation under extreme rainstorm and their further development were explored at a three-year time scale based on unmanned aerial vehicle（UAV） aerial photography and high-precision modern measurement technology. The results showed that: 1）The hillslopes with ephemeral gullies were more conducive to newly formed permanent gullies under the '7.26' extreme rainstorm in northern Shaanxi in 2017. A total of 43 newly formed permanent gullies occurred in the three small watersheds during the rainstorm, of which 67.44 % were connected with ephemeral gullies. 2）The existence of ephemeral gully in the upstream of permanent gully head had an important influence on permanent gullies morphological characteristics. The permanent gullies connected with the ephemeral gullies were significantly longer than that disconnected with the ephemeral gullies, and the width of the permanent gullies was not significantly influenced. 3）Upslope ephemeral gullies also significantly affected the development of permanent gullies connected downslope. On the 3 a time scale after the ' 7.26 ' rainstorm in 2017, the permanent gullies connected with ephemeral gullies were more active in the subsequent development process than the ones not connected with ephemeral gullies. The proportion of gully head development in permanent gullies for newly formed permanent gully heads connected with ephemeral gullies was 1.67 times that of non-ephemeral gully-connected permanent gullies. and the gully headcut retreat rate was 1.80 times. For the existing permanent gullies, the above rates were 3.11 times and 2.17 times, respectively. This study can provide a theoretical basis for more effective management of gully erosion in the hilly-gully region of Loess Plateau. [ABSTRACT FROM AUTHOR]

In recent years, the impact of global warming is becoming increasingly noticeable, and extreme weather is causing rapid and frequent environmental changes. Rice is Taiwan*s most important crop and one of the world's major food crops. Rice lodging majorly affects the livelihoods of farmers, and the government has been attempting to introduce technological tools to help crop damage assessment, so eligible farmers can receive compensation more efficiently. This article introduces a highly efficient assessment method through unmanned aerial vehicle based on multispectral remote sensing and AI machine learning technology on edge computing devices to assess rice lodging. The machine learning model is capable of accurately inferencing rice paddies and lodging areas with the objective and quantifiable features. This assessment method can save significant amounts of time and money for the assessment of lodging damages. This technology has the potential to be applied to large-scale rice paddy management and disaster assessments cost-effectively. [ABSTRACT FROM AUTHOR]

In order to find the spatial-temporal variation and development trends of ecological environment quality in Guanzhong Plain urban agglomeration from 2000 to 2020, the Google Earth Engine(GEE)platform was used to optimize the reconstruction of MODIS images of Guanzhong Plain from June to September during 2000-2020, and remote sensing ecological index(RSEI)was constructed with greenness, humidity, temperature and dryness; the spatial-temporal variation of ecological environmental quality in the region were analyzed; a geodetector model was used to quantify the effects of ten environmental and human factors on the changes of RSEI, and a regression model was constructed by geographic weighted regression(GWR)to analyze the spatial differences of driving factors. The results show that from 2000 to 2020, the ecological environmental quality in Guanzhong Plain urban agglomeration improves year by year; the average RSEI increases from 0.448 to 0.565, with a high distribution in the south and a low distribution in the north. The improved area of ecological environmental quality in Guanzhong Plain urban agglomeration accounts for 70.8%, and the degraded area accounts for 22.9%; the improved area remarkably distributes in Qingyang city, Pingliang city and Tianshui city, and the degraded area mainly distributs in Xi'an city and its surrounding areas. The vegetation type has the greatest impact on the ecological and environmental quality of Guanzhong Plain, with an impact power of up to 0.46. During the study period, the overall ecological environmental quality in Guanzhong Plain urban agglomeration improves with local deterioration in cities with intensive human activities and their surrounding areas. [ABSTRACT FROM AUTHOR]

Studying the joint emission levels of nitrogen dioxide (NO2) and carbon dioxide (CO2) in cities is of great significance for achieving synergistic pollution and carbon reduction, and improving overall efficiency. The anthropogenic emissions of NO2 and CO2 are closely related, and their concentration ratios can reflect the comprehensive level of atmospheric pollutants and carbon emissions in cities. The high-resolution datasets of NO2 vertical column concentration (XNO2) and CO2 vertical column concentration (XCO2) from 2019 to 2020 in China were reconstructed by using TROPOMI, OCO-2, and OCO-3 satellite remote sensing data and environmental auxiliary variables based on XGBoost algorithm, and the concentration ratios of XNO2 annual average and ΔXCO2 (the difference between XCO2 observation and background values) in 31 cities (abbreviated XNO2/ΔXCO2) were analyzed. The results show that the algorithm is superior performance in reconstructing satellite XNO2 and XCO2 with R² of 0.90 and 0.96 for sample-based holdout-validation, respectively. Urban XNO2 is positively correlated with ΔXCO2 (r=0.66) . Cities in advanced development level have higher XNO2/ΔXCO2 compared with cities in averaged development level. Due to the difference in energy utilization efficiency, the XNO2/ΔXCO2 of cities in advanced development level is negatively correlated with urban GDP, permanent population and air pollution level (r=-0.35, -0.38, and -0.10), while that of cities in averaged development level is positively correlated (r=0.37, 0.32, and 0.47). The XNO2/ΔXCO2 is related to the level of urban pollution control and energy utilization, and the results can provide support for urban air pollution control. [ABSTRACT FROM AUTHOR]

Remote sensing technology has become an important tool for monitoring the growth status of potato crops by virtue of its ability to obtain information on the growth of potato crops in a rapid, real-time, accurate, and nondestructive manner. Potato crops have aspects unique to remote sensing monitoring studies because their stems and leaves and tubers grow above and below ground, respectively, and their spectral reflection mechanisms are all different from those of grass crops (such as rice, maize, and wheat). This study summarizes the characteristics of remote sensing technology, the types of remote sensing platforms and remote sensing monitoring methods, and the progress of domestic and foreign research in applying remote sensing technology in the fields of potato crop acreage extraction, disease management, growth monitoring and yield estimation, analyzes the shortcomings of remote sensing technology in monitoring growth information of potato crops, makes further relevant suggestions for this problem, and points out exploring information on spatial distribution of potato crops and changes of potato crops, efficient monitoring of potato crop growth, potato crop yield prediction model, and assimilation mechanism of remote sensing model and agronomic model of potato crops are the future development directions. [ABSTRACT FROM AUTHOR]

Rice shrimp co-cultivation, as a circular agriculture model that integrates ecological, economic, and social benefits, has been widely applied in China. Accurately obtaining information on the planting area, spatial distribution, and changes of rice shrimp co-cultivation is of great significance for ensuring safe production, social supply of rice and shrimp, and formulating corresponding management and planning policies by government departments. Most of the existing methods for extracting rice shrimp farming fields used medium-resolution temporal images as data sources to analyze the characteristics of water or vegetation changes within a year to construct models. The accuracy of the extraction depended on the results of water or vegetation analysis, the boundaries were unclear and there was a lot of noise. To solve the problems of complex temporal analysis, low accuracy of extraction results, and incomplete boundaries in existing extraction methods, this paper used high-resolution images of a single temporal phase as the data source, without analyzing the temporal features of rice shrimp farming fields. Only the spatial features of rice shrimp farming fields on high-resolution images were used for high-precision extraction. We proposed a deep learning semantic segmentation model named EASNet (edge assisted segmentation network), which was mainly composed of three modules: feature extraction module (FE Module), edge assist module (EA Module), and information fusion Module (IF Module). The FE Module was used to extract multi-level and fused contextual semantic information features of the target object. The EA module used the unique edge "shrimp groove" of rice and shrimp farming fields as an auxiliary information for separate segmentation in the designed EA Module. The IF Module integrated the outputs of the FE Module and the EA Module, enabling the main task to not only enhanced the boundary structure information of rice shrimp farming fields but also learned the unique spatial and semantic information of rice shrimp farming fields. The experimental results showed that with the enhancement of the EA Module, the segmentation results of rice shrimp farming fields were more complete and the boundaries were clearer. The IoU (Intersection over Union) of semantic accuracy and the F1-Score of boundary accuracy were improved by 1.5% and 5.8%, respectively. The recall, IoU, and F1-Score of overall semantic accuracy reached 0.970, 0.964, and 0.930, respectively. The Recall and F1-Score of boundary accuracy reached 0.864 and 0.859, while the Recall and F1-Score of relaxed boundary accuracy reached 0.876 and 0.913. The trained EASNet model was applied to the whole area of Xuyi County, and the spatial distribution map of rice shrimp farming fields in Xuyi County in 2020 was obtained. In comparison with the results of rice shrimp breeding fields extracted by the traditional Water Seasonal Difference method and the Random Forest method, our method obtained the optimal results with an OA(overall accuracy) of 96.71% and a Kappa coefficient of 0.934. The method used in this paper has higher accuracy, lower missing rate, more regular boundaries, and fewer instances of similar integration in extracting results. It can provide a basis for natural resource surveys, and government departments to formulate corresponding rice shrimp breeding management and planning policies. [ABSTRACT FROM AUTHOR]

The rapid and accurate extraction of cultivated land is essential for supporting the protection of cultivated land and controlling cultivated land use. With the rapid development of high-resolution remote sensing and artificial intelligence technology, high-resolution cultivated land extraction has gradually transitioned from traditional pixel-based and object-oriented classification algorithms to intelligent cultivated land extraction represented by deep learning. Although many achievements have been made, the new technologies still face significant challenges. First, we sort out and analyze the research status of cultivated land extraction based on traditional machine-learning approaches and deep-learning techniques, which expounds on the necessity in cultivated land extraction. Second, we introduce the basic principle of deep semantic segmentation technology and the experimental process of cultivated land extraction, and summarize the state-of-the-art intelligent cultivated land extraction algorithms. Finally, focusing on some shortcomings of intelligent cultivated land extraction, the development trend of intelligent cultivated land extraction is discussed. [ABSTRACT FROM AUTHOR]

Traditional satellite remote sensing data acquisition, processing, and application models can no longer meet the demands of populism and real-time applications. In order to meet the real-time intelligent service demand for space-based information in the era of Internet of Things, it is of strategic significance to construct a positioning, navigation, timing, remote sensing, communication (PNTRC) integrative space-based information real-time service system that is system-connected, spatio-temporally fused, and service-oriented. This system supports users of all kinds to obtain information, high-precision positioning and timing, and multimedia communication services anytime and anywhere. Wuhan University has deployed the Luojia series of scientific experimental satellites to verify the integrated intelligent remote sensing satellite technology. In order to further promote the construction of the PNTRC system in China, Wuhan University, in cooperation with the Yantai Municipal Government in Shandong Province, has proposed the Oriental Smart Eye intelligent remote sensing satellite constellation plan. This plan aims to achieve innovative service mode transitions in satellite remote sensing, spanning government, industries, and the general public, by utilizing key technologies such as integrated positioning, navigation, timing, remote sensing, communication, satellite constellation network observation, on-orbit processing, and artificial intelligence. It aims to open up the blue ocean market of commercial remote sensing. [ABSTRACT FROM AUTHOR]

In recent years, significant advancements in large language models and visual foundation models in the field of artificial intelligence have attracted scholars' attention to the potential of general artificial intelligence technology in remote sensing. These studies have propelled a new paradigm in the research of large models for remote sensing information processing. Large remote sensing models, also known as pretrained foundation remote sensing models, are a kind of methodology that employs a vast amount of unlabeled remote sensing images to train large-scale deep learning models. The goal is to extract universal feature representations from remote sensing images, thereby enhancing the performance, efficiency, and versatility of remote sensing image analysis tasks. Research on large remote sensing models involves three key factors, including pre-training datasets, model parameters, and pre-training techniques. Among them, pretraining datasets and model parameters can be flexibly expanded with the increase in data and computational resources, while pre-training techniques are critical for improving the performance of large remote sensing models. This review focuses on the pre-training techniques of large remote sensing models and systematically analyzes the existing supervised single-modal pre-trained large remote sensing models, unsupervised single-modal pre-trained large remote sensing models, and visual-text joint multimodal pre-trained large remote sensing models. The conclusion section provides prospects for large remote sensing models in terms of integrating domain knowledge and physical constraint, enhancing data generalization, expanding application scenarios, and reducing data costs. [ABSTRACT FROM AUTHOR]

Objectives: It is of great importance to precisely acquire the spatial distribution and dynamic variations of plastic greenhouses for agricultural development planning, grain assessment, and ecological environment monitoring. High-resolution remote sensing images offer detailed information of plastic greenhouse shapes and boundaries, satisfies the requirements of precision agriculture, and have irreplaceable advantages in conducting precise surveys of plastic greenhouses in crucial regions. Nevertheless, the mapping model of plastic greenhouses meets the challenges including inadequate prior information and unbalance between mapping accuracy and model complexity. Methods: This paper introduces a semantic segmentation model for plastic greenhouse that integrates the advanced plastic greenhouse index (APGI) and high-resolution red green blue (RGB) images. The proposed model primarily comprises the APGI semantic branch, RGB semantic branch, and RGB detail branch, and it incorporates an attention mechanism to effectively merge reliable prior information from the APGI index and spatial detail information derived from high-resolution RGB images. Furthermore, a model training mechanism guided by boundary information is proposed to enhance the precise recognition capability of the proposed model for plastic greenhouse boundary pixels. Results: The study reveals that, within an attention-guided fusion framework, the medium-resolution APGI can significantly enhance the identification capabilities of plastic greenhouses in high-resolution images. The proposed model can not only improve the mapping accuracy of plastic greenhouses but also greatly simpl-ifie the parameter size and computational complexity of the semantic segmentation model. Conclusions: This study meets the demands for rapid and accurate extraction of plastic greenhouse in high-resolution remote sensing images. [ABSTRACT FROM AUTHOR]

In order to expand the application of remote sensing prospecting technology in gold prospecting, this paper takes the Zhuanshanzi area in Inner Mongolia as an example, and selects Landsat 8 OLI remote sensing data. On the basis of image data preprocessing, the regional tectonic information is enhanced by the fusion of B7B5B3 band. The principal component analysis is used to extract the alteration anomaly information through threshold segmentation. Based on the geological characteristics and remote sensing interpretation information of the region and Zhuanshanzi Gold Deposit, the prospecting indicator of gold deposits in Zhuanshanzi area is established, and 3 prospect areas for prospecting prediction are delineated. The field verification shows that there is gold mineralization in the prospect area. It is believed that the linear and ring structure and remote sensing anomaly information extraction based on OLI remote sensing images can be used as auxiliary means for gold prospecting prediction. [ABSTRACT FROM AUTHOR]

Hetao Irrigation District is one of the most important grain and oil production bases in China. The spatial and temporal evolution of the cropping structure over the years can provide the basic data for the yield of farmland and the amount of water for irrigation in the future adjustment of agricultural structures. In this study, the errors from the different sensors were calibrated to calculate the normalized difference vegetation index (NDVI) of the irrigation area for the whole year using Landsat-5 and Landsat-8 remote sensing images. The MODIS NDVI data was also used to interpolate the missing areas. The MODIS NDVI data was interpolated and S-G filtered for the missing areas of the image. The spectral and texture features were also calculated after that. The samples were then augmented to obtain 13004 samples with a uniform distribution. The deep learning of the multilayer perceptron (MLP) neural network was constructed to adjust the MLP model for the actual planting and climatic conditions. The model was then migrated each year to obtain the planting structure from 2000 to 2021. Finally, a systematic analysis was made of the multi-year spatial and temporal changes. The results showed that the overall classification accuracy reached 89% in 2021, and the Kappa coefficient was 0.87. Overall, the classification accuracy was ranked in descending order as follows: wheat, maize, alfalfa, vegetables, sunflower (melons), and rice. Specifically, the accuracies of the rice and melons were lower at 0.86 and 0.88, respectively, in terms of the user’s accuracy indexes. All accuracies were higher than 90% in the main crops of maize, wheat, and alfalfa. The relative error of the area in each year was less than 8%, compared with the statistical planting area. The smaller relative error was found in the sunflower, corn and wheat, whereas, the larger one was vegetables and melons (5% to 8%). A comparison of the MLP classification revealed that there was a relatively high consistency of the spatial distribution. Meanwhile, the contiguous planting area was distributed in the form of plots. The planting area of sunflower and maize showed an upward trend between 2000-2021, whereas, the planting area of wheat shrunk considerably. The sunflower was distributed mainly in the eastern region in 2015-2021, maize growing areas are mainly distributed in the central and western regions, but in 2010 scattered distribution, no contiguous planting area. The wheat planted area accounted for a relatively large proportion of the total in 2000-2010, mainly in the northwestern region and eastern part of the irrigation area, and then shrunk considerably in 2010. The evolution of sunflower acreage occurred more during the study period, followed by maize, all of which was mainly transferred from the wheat. The high classification accuracy and migration of the improved model can provide a strong reference for agricultural management and rational resource utilization in the Hetao Irrigation District of Inner Mongolia. [ABSTRACT FROM AUTHOR]

Satellite remote sensing monitoring has advantages in compensating for insufficient ground measurements, comprehensively understanding behavior changes in water conservancy projects and identifying risk sources. This paper examines optical and radar remote sensing monitoring technologies, outlining their current application in water conservancy projects. It highlights challenges in image classification, recognition, InSAR deformation measurement, and their integration with risk assessment. In response to the application requirements for the construction progress and operation management of projects, the business application path of remote sensing monitoring for water conservancy projects is discussed from three typical application scenarios: progress supervision of projects under construction, precise deformation monitoring of operational projects, and risk identification and early warning. It analyzes technical bottlenecks, such as quantification, spatiotemporal resolution, and multi-scale collaborative monitoring. Finally, it outlines future trend in remote sensing monitoring technology for water conservancy projects. [ABSTRACT FROM AUTHOR]

River bathymetry (RB) is a fundamental dataset in the field of river research. However, mountainous regions often lack comprehensive data due to topographical and transportation challenges. Remote sensing technology provides an innovative method for estimating RB. In this study, the theoretical relationship between the water level and the river width is established by generalizing the channel cross-section shape. A novel RB estimation method was proposed, integrating the Hydroweb dataset and Sentinel-1 images. The impacts of exposure, reach-average length, and remote sensing observation errors on estimation accuracy were systematically analyzed. The method was applied to the Upper Yangtze River to evaluate its potential for estimating river discharge. Results reveal that: ① The estimation error of the riverbed elevation ranges from 4.00 m to 4.06 m, with the estimated cross-section representing 73.69% to 80.29% of the actual area, indicating precise RB estimation. ② Exposure rate emerges as a primary factor, significantly enhancing estimation accuracy. An appropriate reach-average length improves the estimation precision and optimal length of 10km is advised for the Upper Yangtze River. Furthermore, the accuracy of RB estimation is more susceptible to water level errors in remote sensing than to river width. ③ The method demonstrates the potential to estimate river discharge achieving a Nash efficiency coefficient of 0.92. The research outcome can provide a novel approach to RB monitoring in data-scarce regions. [ABSTRACT FROM AUTHOR]

Rapid extraction and precise identification of freshwater aquaculture ponds can play an important role in the management and decision-making of the aquaculture industry. Satellite remote sensing can be expected to provide the critical extraction of freshwater aquaculture ponds, due to its rapid, timely and effective way of Earth observation. However, it is still elusive to delineate the surface area and the variation in the aquaculture ponds from the satellite remote sensing images. In this study, an improved classification was introduced to delineate the changes in aquaculture ponds from Landsat remote sensing images using the coordinate attention with the U-Net neural network model. Firstly, the dataset was collected from the freshwater aquaculture ponds in the Gaochun district of Nanjing City, Jiangsu Province, China. Landsat remote sensing images were also captured from 1985 to 2021, and supplemented by GF-1, GF-2 satellite data and field survey data. Secondly, a coordinate attention model with U-Net as the backbone was created to fully extract the spatial information of features, where the information remained on the fragmented aquaculture ponds. Finally, the long-term Landsat images were selected to explore the temporal and spatial changes of freshwater aquaculture ponds. Six state-of-the-art models were also utilized to verify the improved model. The experimental results show that the performance of the improved model outperformed the rest, in terms of extraction accuracies. Furthermore, there was a dramatic variation in the surface area and distribution of freshwater aquaculture ponds in the study period. Specifically, the surface areas of aquaculture ponds increased from 0.48 to 36.92 km² in the early stages from 1985 to 2000, with an annual increase of 2.43km². There was rapid growth in the second stage from 2000 to 2017. Among them, the surface areas of aquaculture ponds increased from 36.92 to 234.47 km², with an annual increase of 11.62 km². A gradual shrinkage was found in the surface areas of freshwater aquaculture ponds from 234.47 to 209.58 km² in the third stage from 2017 to 2021, with an annual decrease of 6.22 km². Moreover, the social and economic factors can be attributed to the main driving factors for the dramatic variation in the surface area of the aquaculture areas. In summary, the high overall accuracies of the improved model were achieved to rapidly extract the aquaculture ponds. Additionally, the food demand and economic value can be the significant driving factors for the rapid changes in the aquaculture ponds, while the policies and human activities can be the key factors for the area changes of the aquaculture ponds. The finding can also provide technical support for the scientific management and decision-making of the aquaculture industry. [ABSTRACT FROM AUTHOR]

Plastic mulch has been widely used to increase the soil temperature with the constant moisture for the high crop yields in the arid regions in northwestern China. Particularly, there has also the application in maize, wheat, and cotton cultivation in Hexi Corridor since the last century. Nevertheless, the mulch residues have significantly damaged the soil structure, leading to environmental degradation. Therefore, it is very necessary for the spatial and temporal distribution of plastic mulch, in order to quantify the impact on the local environment. However, the wide region and the complex interlocking surface types have posed a great challenge to the spatial and temporal distribution of plastic mulched cropland (PMC) from the field work. Fortunately, large-scale ground objects can be extracted with the development of remote sensing. However it is still lacking in the acquisition and processing of remote sensing data. This study aims to extract the spatiotemporal PMC variations in Hexi Corridor from 2013 to 2022. The multi-source remote sensing data was first collected using Google Earth Engine, including Sentinel-2, Landsat 8, Landsat 7, and MODIS. A time window of PMC extraction was also proposed using the variance of the Normalized Difference Vegetation Index (NDVI) within one year. Then, the Timely Plastic-mulched cropland Extraction Method (TPEM) was applied to the complex and variable ground objects for the extraction of the PMC area. About 500 sample points were randomly selected and then labelled as the PMC and non-PMC for the accuracy assessment. The Overall Accuracy (OA), User Accuracy (UA), Quantity Disagreement (QD), and Allocation Disagreement (AD) were also used to evaluate the performance of TPEM. Finally, the spatiotemporal pattern of PMC was analyzed to explore the potential controls on the interannual variation in the PMC area using Water Yearbook Bulletin data. The results showed that: 1) The best time window was the first half of April for the PMC extraction, which also coincided with the sowing of major local crops. The half month before and after this period was also beneficial to extracting the PMC information. 2) The TPEM reached 80% overall accuracy with the QD less than 4%. The error was attributed to the disagreement between the sample points and the real PMC area. 3) The average PMC area was about 383 000 hm², of which about 55.6% was in the Heihe Basin. The PMC was located near the inland river systems. Approximately 41.9% of the cropland used plastic mulch, while the PMC area decreased by 32.4% in the last decade. Meanwhile, there was a 46% increase in the water-saving irrigation area, whereas, there was a 14.9% decrease in the irrigation water from 2013 to 2019, indicating the significant correlation with the decrease in the PMC area. In conclusion, the application of plastic mulch was ever decreasing with the development of water-saving irrigation. This finding can be an essential reference for the survey of agricultural resources in the Northwest Arid Region of China. [ABSTRACT FROM AUTHOR]

Accurate and timely nitrogen status diagnosis is crucial for the nitrogen application management and yield prediction of winter rapeseed. Traditional, destructive manual measurement cannot fully meet the large-scale production in recent years, due to the tedious, time-consuming and laborious. Fortunately, the non-destructive and efficient hyperspectral technology can serve as the necessary means in modern precision agriculture. This study aims to realize the rapid and accurate diagnosis for the prediction of the nitrogen nutrition status of winter rapeseed. Data sources were collected from 90 plant samples (above-ground biomass and plant nitrogen concentration) and hyperspectral measured data at the bud stage of winter oilseed rape under different mulching and nitrogen application treatments for two consecutive years (2022-2023). Eight typical spectral indices (16 in total) were first calculated using the correlation coefficients between the original and the first-order differential (FD) spectrum and the nitrogen nutrition index (NNI). Secondly, the correlation matrix was used to extract the best spectral combination. The optimal spectral index was then selected from the correlation coefficient with NNI. Finally, the optimal spectral index was divided into three groups of input variables for the model (combination 1: the five original spectral indices; combination 2: the five first-order differential spectral indices; combination 3: the five spectral indices, with the highest correlation coefficient with NNI). Support vector machine (SVM), random forest (RF), extreme learning machine (ELM), and back propagation neural network (BPNN) were used to construct the NNI estimation model of winter oilseed rape. Determination coefficient (R² ), root mean square error (RMSE), and mean relative error (MRE) were used to evaluate the accuracy of the model. The results show that the correlation coefficient between the first-order differential spectral index and NNI was greater than that of the original, indicating a better NNI prediction using the first-order differential spectral index. There were high correlation coefficients between the spectral indices selected by the three combinations and NNI. The wavelength combinations were located in the red edge (670-760 nm), indicating the NNI variation representing the characteristic information in the red edge. The FDSAVI spectral index shared the highest correlation coefficient (0.674) with NNI. Among them, the wavelength combination was located at 712 and 678 nm. An optimal combination of the RF model and combination 2 were achieved to combine the optimal input variables of the NNI estimation model. The R² value, RMSE, and MRE on the validation set were 0.823, 0.079, and 7.513%, respectively, indicating the high accuracy of the optimal model. The finding can provide a technical basis for the potential application of remote sensing for the monitoring and diagnosis of plant nitrogen nutrition. [ABSTRACT FROM AUTHOR]

Objectives: It is difficult for the traditional traffic information acquisition method to obtain the real-time dynamic traffic flow information of large scope and full coverage. With urbanization, the increasing number of supertall buildings makes the city skyline a favorable platform for earth observation. This paper studies the urban remote sensing video production and traffic information extraction method using the urban skyline as the observation platform. Methods: First, the earth observation data shooting is carried out in the super-tall buildings. Then, the original oblique observation data is corrected by orthography, which is further fused with the satellite image to generate a large range of remote sensing video data of the city. Finally, we train a deep learning model and use it for vehicle classification and identification. The number and density of vehicles in the area are calculated based on the identification results. Results: This paper carries out data collection and processing analysis on the sightseeing floor of Ping, an Building in Shenzhen. Results show that the proposed method can produce low-cost, long-time, large-scale and high-quality urban remote sensing video. The vehicle detection based on the remote sensing video have high accuracy. Conclusions: The proposed framework can be used to effectively monitor the regional traffic flow and sever for the smart city management. Furthermore, based on its capability of long-time, high-resolution, full-coverage and real-time remote sensing, this method can also be applied to many other urban management fields, such as urban disaster emergency response, crowd monitoring, and construction site monitoring. [ABSTRACT FROM AUTHOR]

Objectives: Vegetation plays an important role in ecological environment monitoring. Studying vegetation cover change can provide reference for regional ecological environment protection. The region of Altay in Xinjiang belongs to the ecological function zone of water conservation mountain grassland, with rich natural resources and beautiful scenery. However， the ecological environmental problems were gradually emerging under the national development strategies. Therefore, monitoring the dynamics of vegetation change in the area with multi-source remote sensing data is necessary to explore the relationship between the economic development and ecological environment protection. Methods: This paper uses multisource remote sensing long time series data with breaks for additive seasonal and trend, topographical position and geostatistical analysis methods to monitor the vegetation cover dynamics in Altay of Xinjiang, China during 2000—2019. Results: This paper processed large number of multisource remote sensing data along with human activity data and showed that during the study period, the number of vegetation breakpoints increased year by year from 2003 to 2009, and then gradually decreased after 2009. Meanwhile, seven types of vegetation breakpoint trends were detected, among them, the more number of breakpoints from degradation to growth type was identified, and the types of disturbance degradation and continuous degradation were less. From the interannual change curve, the dynamic change of vegetation coverage showed a trend of first decreasing and then increasing, that is, the vegetation coverage showed a degradation trend from 2000 to 2008 and a significant improvement trend from 2008 to 2019, and the degradation was greater than the improvement. Meanwhile, the east slope with an elevation of more than 900 m and a slope of more than 15° (from north to east and from south to east) is the dominant topographic location of vegetation degradation type, accounting for 62.4%. Conclusions: The protection of ecological environment in Altay of Xinjiang still needs to be further strengthened and effective protection measures should be taken. [ABSTRACT FROM AUTHOR]

To solve the problem that the existing methods were not good at detecting vehicles on ground roads, a vehicle detection method was proposed based on deep learning and high-resolution remote sensing images. The interested regions of high-resolution remote sensing images were extracted, and the multi-object detection of vehicles was performed by the improved YOLOv3 model, which utilized multi- scale features for object detection. The training process was determined and carried out on the high- resolution remote sensing image dataset, and the image recognition and detection tests were performed with accuracy rate, recall rate and F value as evaluation indicators. The results show that using the proposed method to analyze the high-resolution remote sensing images, the accuracy rate, recall rate and F value are 98.01%, 97. 23% and 97. 57%, respectively, and the traffic flow per second can be obtained according to the statistics of the detection results. The proposed method can be used as effective supplementary method for monitoring the distribution of information on ground vehicles. [ABSTRACT FROM AUTHOR]

To solve the problem that the existing methods were not good at detecting vehicles on ground roads, a vehicle detection method was proposed based on deep learning and high-resolution remote sensing images. The interested regions of high-resolution remote sensing images were extracted, and the multi-object detection of vehicles was performed by the improved YOLOv3 model, which utilized multi-scale features for object detection. The training process was determined and carried out on the high-resolution remote sensing image dataset, and the image recognition and detection tests were performed with accuracy rate, recall rate and F value as evaluation indicators. The results show that using the proposed method to analyze the high-resolution remote sensing images, the accuracy rate, recall rate and F value are 98.01%, 97.23% and 97.57%, respectively, and the traffic flow per second can be obtained according to the statistics of the detection results. The proposed method can be used as effective supplementary method for monitoring the distribution of information on ground vehicles. [ABSTRACT FROM AUTHOR]