Search

The timely and accurate acquisition of information on the distribution of the crop planting structure in the Loess Plateau of eastern Gansu Province, one of the most important agricultural areas in Western China, is crucial for promoting fine management of agriculture and ensuring food security. This study uses multi-temporal high-resolution remote sensing images to determine optimal segmentation scales for various crops, employing the estimation of scale parameter 2 (ESP2) tool and the Ratio of Mean Absolute Deviation to Standard Deviation (RMAS) model. The Canny edge detection algorithm is then applied for multi-scale image segmentation. By incorporating crop phenological factors and using the L1-regularized logistic regression model, we optimized 39 spatial feature factors—including spectral, textural, geometric, and index features. Within a multi-level classification framework, the Random Forest (RF) classifier and Convolutional Neural Network (CNN) model are used to classify the cropping patterns in four test areas based on the multi-scale segmented images. The results indicate that integrating the Canny edge detection algorithm with the optimal segmentation scales calculated using the ESP2 tool and RMAS model produces crop parcels with more complete boundaries and better separability. Additionally, optimizing spatial features using the L1-regularized logistic regression model, combined with phenological information, enhances classification accuracy. Within the OBIC framework, the RF classifier achieves higher accuracy in classifying cropping patterns. The overall classification accuracies for the four test areas are 91.93%, 94.92%, 89.37%, and 90.68%, respectively. This paper introduced crop phenological factors, effectively improving the extraction precision of the shattered agricultural planting structure in the Loess Plateau of eastern Gansu Province. Its findings have important application value in crop monitoring, management, food security and other related fields. [ABSTRACT FROM AUTHOR]

Thunderstorm gusts are a common form of severe convective weather in the warm season in North China, and it is of great importance to correctly forecast them. At present, the forecasting of thunderstorm gusts is mainly based on traditional subjective methods, which fails to achieve high-resolution and high-frequency gridded forecasts based on multiple observation sources. In this paper, we propose a deep learning method called Thunderstorm Gusts TransU-net (TG-TransUnet) to forecast thunderstorm gusts in North China based on multi-source gridded product data from the Institute of Urban Meteorology (IUM) with a lead time of 1 to 6 h. To determine the specific range of thunderstorm gusts, we combine three meteorological variables: radar reflectivity factor, lightning location, and 1-h maximum instantaneous wind speed from automatic weather stations (AWSs), and obtain a reasonable ground truth of thunderstorm gusts. Then, we transform the forecasting problem into an image-to-image problem in deep learning under the TG-TransUnet architecture, which is based on convolutional neural networks and a transformer. The analysis and forecast data of the enriched multi-source gridded comprehensive forecasting system for the period 2021–23 are then used as training, validation, and testing datasets. Finally, the performance of TG-TransUnet is compared with other methods. The results show that TG-TransUnet has the best prediction results at 1–6 h. The IUM is currently using this model to support the forecasting of thunderstorm gusts in North China. [ABSTRACT FROM AUTHOR]

The coronavirus first appeared in China in 2019, and the World Health Organization (WHO) named it COVID-19. Then WHO announced this illness as a worldwide pandemic in March 2020. The number of cases, infections, and fatalities varied considerably worldwide. Because the main characteristic of COVID-19 is its rapid spread, doctors and specialists generally use PCR tests to detect the COVID-19 virus. As an alternative to PCR, X-ray images can help diagnose illness using artificial intelligence (AI). In medicine, AI is commonly employed. Convolutional neural networks (CNN) and deep learning models make it simple to extract information from images. Several options exist when creating a deep CNN. The possibilities include network depth, layer count, layer type, and parameters. In this paper, a novel Xception-based neural network is discovered using the genetic algorithm (GA). GA finds better alternative networks and parameters during iterations. The best network discovered with GA is tested on a COVID-19 X-ray image dataset. The results are compared with other networks and the results of papers in the literature. The novel network of this paper gives more successful results. The accuracy results are 0.996, 0.989, and 0.924 for two-class, three-class, and four-class datasets, respectively. [ABSTRACT FROM AUTHOR]

Automatic weed identification and detection are crucial for precision weeding operations. In recent years, deep learning (DL) has gained widespread attention for its potential in crop weed identification. This paper provides a review of the current research status and development trends of weed identification in crop fields based on DL. Through an analysis of relevant literature from both within and outside of China, the author summarizes the development history, research progress, and identification and detection methods of DL-based weed identification technology. Emphasis is placed on data sources and DL models applied to different technical tasks. Additionally, the paper discusses the challenges of time-consuming and laborious dataset preparation, poor generality, unbalanced data categories, and low accuracy of field identification in DL for weed identification. Corresponding solutions are proposed to provide a reference for future research directions in weed identification. [ABSTRACT FROM AUTHOR]

Due to the periodic changes of climate, cracks are widely developed in expansive soils, leading to the destruction of soil integrity, the deterioration of physical strength, and eventually the instability of the expansive soil slope and other disasters. In this paper, a deep learning-based modeling method was proposed for soil crack networks characterization through steps of segmentation, quantification and simulation. Inspired by the U-Net convolutional neural network, a dilated convolution module was added to the backbone to enhance the crack segmentation capability and a subpixel edge detection algorithm was followed for accurate crack edge detection. Then, a deterministic or stochastic method was designed for crack network simulation. A case study of an expansive soil slope on the bank of Wadong Main Canal in the Pi-Shi-Hang Irrigation District, China, was conducted. Results show that the dilated U-Net model gained 0.24 and 0.25 improvement in F1-score and IoU (Intersection over Union) comparing to the conventional segmentation method (Otsu) and crack edge precision was further improved by 5.38%. The performances of proposed method including the image labeling, effect of crack thickness and environmental conditions, etc., are also explored. To validate the simulated crack network, crack areas at different depths were measured through X-ray images. Comparing with the simulation result, mean error rate of 1.05% was achieved. Thus, the proposed method can efficiently characterize soil crack networks using the field acquired crack images. [ABSTRACT FROM AUTHOR]

Chinese calligraphy is an important part of Chinese national culture and art and part of the essence of Chinese national culture. Stone calligraphy is one of the important elements of Chinese calligraphy art. Stone carved calligraphy characters have high cultural and artistic value. Therefore, accurately recognizing stone carved calligraphy characters are of great importance. Stone carved calligraphy can identify hard-to-preserve stone calligraphy paper materials in electronic data that can be preserved for a long time, thereby offering important reference materials for the study of the historical development of Chinese calligraphy art. Moreover, with the development of science and technology and the investment of China in cultural and artistic undertakings, computer-aided calligraphy character recognition technology is also constantly improving, and its application in calligraphy recognition is becoming increasingly extensive. This article aims to study a method of stone inscription calligraphy recognition based on convolutional neural networks. In this paper, we use an image recognition and optimization method consisting of a convolutional neural network to carry out an experiment with stone inscription calligraphy characters. It was concluded that the recognition accuracy of stone calligraphy characters by the convolutional neural network reached 99.2%, indicating that this stone calligraphy character recognition method based on a convolutional neural network has a good ability to recognize stone calligraphy characters. [ABSTRACT FROM AUTHOR]

Landslide susceptibility mapping (LSM) is essential for the spatial prediction of landslides and risk prevention. Physically based LSM models are confined by oversimplifications of physical processes and limited information about soil properties. Data-driven LSM models may give reliable results only when the training and the testing data have high similarity, and application in regions with different geological conditions is often inapplicable. This paper proposes four hybrid data-driven and physical models and compares these models in terms of cross-regional generalization ability and prediction uncertainty. The effects of physical module performance on the hybrid model are analyzed. For the physical modules of the four hybrid models, two-dimensional (2D) physically based models, TRIGRS and the infinite-slope stability models (ISSMs), and the three-dimensional (3D) physically based model, Scoops3D, are adopted. The data-driven modules all adopt the convolutional neural network (CNN) model. Two towns in the Three-Gorge Reservoir area of China are used as the training and testing areas. The results show that all hybrid models have better generalization ability than using the data-driven module exclusively. The prediction uncertainty is significantly reduced by pre-selecting training samples using the physical module. The optimal hybrid model is the one that integrates CNN and ISSM (under the saturated condition). It is then applied to a new region (Wushan County) to further validate the generalization ability. It can make accurate predictions without calibrating the trained model using new data from the validation area. Finally, the effectiveness of model averaging for improving the prediction performance is verified. Using model averaging, the AUC value in the validation area yields 0.834, which is even higher than the original realizations, with AUC values ranging from 0.683 to 0.817. Therefore, the optimal hybrid model can be directly used for LSM in the Three-Gorge Reservoir, and the findings can provide valuable guidance for generalization ability improvement and prediction uncertainty reduction of LSM models in other countries and regions. [ABSTRACT FROM AUTHOR]

Simple Summary: Most management methods for poultry farming currently rely on human labor. Such management is labor-intensive and inefficient, especially in identifying poultry growth stages. Given the lack of a high-precision artificial method for chick's day-age detection, a high-performance day-age classification and detection model for chicks was proposed based on artificial intelligent techniques. This method can detect and classify chicks in six different living stages from 1 to 32 day-ages, and the accuracy is 95.2%, superior to other current ones. In order to apply this method in practical scenarios, it has been deployed into an application based on the IOS system, which can recoganize the day-age of chicks by capturing real-time photos. The system is currently deployed in Rizhao City, Shandong Province, China. It helps chicken farm staff automatically detect the behavior of chickens, whose excellent working effect proves the robust availibility of the proposed method. Thanks to the boom of computer vision techniques and artificial intelligence algorithms, it is more available to achieve artificial rearing for animals in real production scenarios. Improving the accuracy of chicken day-age detection is one of the instances, which is of great importance for chicken rearing. To solve this problem, we proposed an attention encoder structure to extract chicken image features, trying to improve the detection accuracy. To cope with the imbalance of the dataset, various data enhancement schemes such as Cutout, CutMix, and MixUp were proposed to verify the effectiveness of the proposed attention encoder. This paper put the structure into various mainstream CNN networks for comparison and multiple ablation experiments. The final experimental results show that by applying the attention encoder structure, ResNet-50 can improve the accuracy of chicken age detection to 95.2%. Finally, this paper also designed a complete image acquisition system for chicken houses and a detection application configured for mobile devices. [ABSTRACT FROM AUTHOR]

Groundwater level (GWL) prediction is important for ecological protection and resource utilization; it helps in formulating policies for artificial groundwater recharge, modifying the number of extraction wells, etc., and can support sustainable human development as well as inform water resource management decisions. However, climate change, anthropogenic impacts, and the complex coupling between surface water and groundwater increase the difficulty of predicting groundwater levels. The model proposed in this paper combines external data as well as multiple models. The method leverages long and short-term memory (LSTM) and convolutional neural network (CNN) models, combined with secondary modal decomposition and slime mould algorithm (SMA), together with an adaptive weight module (AWM). The study applies this method to predict GWL for three different hydrological conditions in China, specifically for the Jinan Baotu Spring, Heihu Spring, and Zhongtianshe watershed of Taihu Lake. A comparison of metrics such as mean absolute error and Nash efficiency coefficient for single and hybrid models shows that the model in this paper is more advantageous than the single model and other hybrid models. The interpretability of the model is enhanced by SHAP values that demonstrate the degree of contribution of the input variables. This paper uses SHAP analyses to identify the key drivers affecting groundwater levels. These factors must be detected in order to develop groundwater resource protection measures. [Display omitted] • Multivariate fusion data including hydrology and meteorology are used as model input. • A secondary modal decomposition module for historical groundwater level data was utilized. • The neural network hyperparameters are optimized using the slime mould algorithm. • Aggregate subsets of prediction with adaptive modules rather than linear summation. • The interpretable SHAP model measures the degree of influence of external variables. [ABSTRACT FROM AUTHOR]

Passenger flow is an important benchmark for measuring tourism benefits, and accurate tourism passenger flow prediction is of great significance to the government and related tourism enterprises and can promote the sustainable development of China's tourism industry. For daily passenger flow time series data, a passenger flow forecasting method based on convolutional neural network (CNN) and improved quantile regression long short-term memory network (QRLSTM), denoted as CNN-IQRLSTM, is proposed with reconstructed correlation features and in the form of sliding windows as inputs. First, four discrete variables such as whether the day is a weekend and holiday are created by time; then, a sliding window of width 42 is used to pass the passenger flow data into the network sequentially; finally, the loss function of the sparse Laplacian improved QRLSTM is introduced for passenger flow prediction, and the point prediction and interval prediction results under different quartiles are obtained. The application of quantile regression captures the overall picture of the data, enhances the robustness, fit, predictive power and nonlinear processing capability of neural networks, and fills the gap between quantile regression and neural network methods in the field of passenger flow prediction. CNN can effectively handle complex input data, and the improved nonlinear QR model can provide passenger flow quantile prediction information. The method is applied to the tourism traffic prediction of four 5A scenic spots in Jilin Province, and the effectiveness of the method is verified. The results show that the method proposed in this paper fits best in point prediction and has higher prediction accuracy. The MAPE of the Changbai Mountain dataset was 0.07, the MAPE of the puppet palace museum dataset was 0.05, the fit of the Sculpture Park dataset reached 93%, and the fit of the net moon lake dataset was as high as 99%. Meanwhile, the interval prediction results show that the method has a larger interval coverage as well as a smaller interval average width, which improves the prediction efficiency. In 95% of the interval predictions, the interval coverage of Changbai Mountain data is 99% and the interval average width is 0.49. It is a good reference value for the management of different scenic spots. [ABSTRACT FROM AUTHOR]

A hybrid framework by integrating stacking ensemble with two deep learning methods of convolutional neural network (CNN) and recurrent neural network (RNN) is introduced in this paper for landslide spatial prediction in the Three Gorges Reservoir area, China. The proposed framework is summarized in following steps. First, a spatial database consists of 20 landslide conditioning factors and 196 landslide polygons was established. Then, landslide and non-landslide pixels were randomly divided into training (70% of the total) and test (30%) sets. Next, a stacking ensemble method that integrates CNN and RNN was constructed using the training set. Finally, the proposed stacking framework was applied for landslide susceptibility mapping and evaluated. Experimental results demonstrated that the proposed framework can obtain the best predictive capability (0.918) than CNN (0.904), RNN (0.900) and logistic regression (0.877) in terms of area under the receiver operating characteristic curve (AUC). Therefore, it can be useful for landslide disaster management and assessment. [ABSTRACT FROM AUTHOR]

In the past decade, the number of cars in China has significantly raised, but the traffic jam spree problem has brought great inconvenience to people's travel. Accurate and efficient traffic flow prediction, as the core of Intelligent Traffic System (ITS), can effectively solve the problems of traffic travel and management. The existing short-term traffic flow prediction researches mainly use the shallow model method, so they cannot fully reflect the traffic flow characteristics. Therefore, this paper proposed a short-term traffic flow prediction method based on one-dimensional convolution neural network and long short-term memory (1DCNN-LSTM). The spatial information in traffic data is obtained by 1DCNN, and then the time information in traffic data is obtained by LSTM. After that, the space-time features of the traffic flow are used as regression predictions, which are input into the Fully-Connected Layer. In the end, the corresponding prediction results of the current input are calculated. In the past, most of the researches are based on survey data or virtual data, lacking authenticity. In this paper, real data will be used for research. The data are provided by OpenITS open data platform. Finally, the proposed method is compared with other road forecasting models. The results show that the structure of 1DCNN-LSTM can further improve the prediction accuracy. [ABSTRACT FROM AUTHOR]

China is the major producer and consumer of pork all around the world. In China, pork production has accounted for more than 60% of meat production for a long time. Nowadays, the pig industry has gradually shifted from decentralized and extensive breeding to intensive large-scale breeding, promoting production efficiency. At present, as people's demand for the quantity and quality of pork is increasing, China's pig industry faces the problem that the output and quality of pork can not meet people's daily needs. With the rise of artificial intelligence technology in recent years, deep learning technology has been developed rapidly and widely used in image and audio recognition, natural language processing, robotics, bioinformatics, chemistry, finance, and other fields. It is also an essential tool in developing precision livestock farming. The body condition, behavior and health status of pigs could directly affect the income level of pig farms, so through using deep learning technology, we can quickly and accurately acquire the relevant information about pigs, and carry out precise management to improve the feeding efficiency and welfare levels of pigs. This paper expounds on the research progress and application status of deep learning technology used in target pig detection, pig image segmentation, pig body condition and abnormal monitoring, as well as pig behavior recognition. Then we put forward the improvement strategy of deep learning technology used in the process of pig feeding, which make it easier for researchers to understand. At the same time, we summarize and analyze the data sources and datasets, application scope, and model optimization of previous works using deep learning technology in pig breeding. In the field of data sources and datasets, mobile computer vision systems are currently more suitable than systems with many fixed cameras when they are applied in pig houses, therefore, further research could focus on how to use deep learning technology to mobile computer vision systems. Deep learning technology requires a large amount of data to learn data features, which is a massive disadvantage for pig applications. To get pre-training weights suitable for agricultural data sets, a large number of public datasets and a unified dataset standard suited to the pig field should be established. In terms of application scope, due to the short application time of deep learning in the process of raising pigs, many critical occasions are not involved or seldom involved in existing research, so the application scope of deep learning should further expand. In model optimization, optimizing the model to meet the needs of practical application scenarios is the direction of future research. Optimizing the model to achieve a better balance between size and model performance requires further study. Optimizing the model to locate the start and end time of the behavior in unclipped video and recognize the behavior of target pigs is the challenge in identifying the behavior of pigs in future. China's research on deep learning is at the top level, so the application prospect of deep learning in pig farming is highly expected. Combined with the actual production scene of pigs, deep learning technology will significantly contribute to the development of precision livestock farming. [ABSTRACT FROM AUTHOR]

The improvement of photovoltaic (PV) power prediction precision plays a crucial role in the new energy consumption. This paper proposes a composite prediction framework (DC (DWT-DAE)-CNN) consisting of dual clustering and convolutional neural network to achieve day-ahead prediction of PV power. To avoid the temporal uncertainty of PV power and the high-dimensional complexity of numerical weather prediction, the raw data are processed by Discrete Wavelet Transform (DWT) and Deep Autoencoder respectively (DAE) to reduce the data redundancy. Secondly, a dual clustering pattern based on dynamic time warping distance clustering and Fuzzy C-Mean (FCM) clustering is proposed to progressively realize the dynamic characteristics of the power curve and numerical clustering of the weather information data. Finally, the data from PV plants in northeast China are used for validation. The results show that the annual average day-ahead prediction AR of the DC (DWT-DAE)-CNN model can reach 90.17%, which is better than other competing models. In addition, the dual clustering pattern performs better than other traditional clustering patterns with the same predictor. Using this method to predict the PV output power can provide better theoretical guidance for the stable and safe operation of grid-connected PV. • A dual clustering model based on numerical variation and curve morphology variation is proposed. • The power data and NWP data were processed separately using Discrete Wavelet Transform and Deep Autoencoder. • Building multiple convolutional neural network predictors for PV power day-ahead prediction. • The proposed model was shown to outperform the comparison model under different seasons. [ABSTRACT FROM AUTHOR]

The transformation of resource-exhausted urban land is an urgent problem for sustainable urban development in the world today. Obtaining the urban land use type and analyzing the changes in their land use can lead to better management of the relationship between economic development and resource utilization. In this paper, a residual-intelligent module network was proposed to solve the problems of low classification accuracy and missing objects edge information in traditional computer classification methods. The classification of four Landsat-TM/OLI images from 1993–2020 for Jiaozuo city (the first batch of resource-exhausted cities in China) was realized by this method. The results (overall accuracy was 98.61%, in 2020 images) were better than the comparison models (support vector machine, 2D-convolutional neural network, hybrid convolution networks; overall accuracy was 87.12%, 96.16%, 98.46%, respectively) and effectively reduced the loss of information on the edge of the ground objects. On this basis, six main land use types were constructed by combining field surveys and other methods. The characteristics and driving forces of spatial-temporal change in land use were explored from the aspect of social, economic and policy factors. The results showed that from 1993 to 2020 the cultivated land, forest land, water body and other land types in the study area decreased by 690.97 km2, 57.54 km2, 47.04 km2 and 59.43 km2, respectively. The construction land and bare land increased by 839.38 km2 and 15.57 km2, respectively. The transfer of land use types was mainly from cultivated land to construction land, with a cumulative conversion of 920.95 km2 within 27 years. The driving forces of land use in the study area were analyzed by principal component analysis (PCA) and regression analysis. The spatial-temporal evolution of land use types was affected by policy changes, the level of social development and the adjustment in the economy, industry and agriculture structure. The investment in fixed assets and per capita net income in rural areas were the top two influencing factors and their cumulative contribution rate was 94.62%. The findings of this study can provide scientific reference and theoretical support for land use planning, land reclamation in mining areas, ecological protection and sustainable development in Jiaozuo and other resource-exhausted cities in the world. [ABSTRACT FROM AUTHOR]

Estimating the total wind power output from the meteorological information at a province level (called Provincial Regional Wind Power Conversion Model, PRWPCM) plays vital and fundamental roles in energy modeling community and regional wind power forecasting. How to construct a reliable PRWPCM is a real challenge, since PRWPCM involves a large number of widely distributed wind turbines, massive meteorological data across the whole province, and complex nonlinear correlations. This paper proposes a lightweight PRWPCM by integrating Geographic Information System (GIS) analysis technology and Convolutional Neural Network (CNN). First, we conduct the land suitability analysis for wind turbine sites through the multi-criteria GIS layer overlay method to make the provincial wind turbine land suitability map (WTLSM) with scored divisions from the least suitable to the most suitable areas. On this basis, a new fusion mechanism for geographic and meteorological information is proposed, through which the raw meteorological data matrix can be reconstructed to filter and amplify the meteorological information that is more relevant to the total wind power output of the province, and avoid the time-consuming and labor-intensive data collection and processing, large-size model construction and validation. Second, a CNN-based regression architecture is designed to further capture the mapping relationship between the reconstructed meteorological data and total wind power output of the province; each type of meteorological factor is considered as an input channel and the attention modules are introduced to adaptively enhance useful channels and suppress less useful ones. Numerical experiments based on the wind power operation data of Yunnan Province, China, are conducted to validate the superiority of the proposed PRWPCM via benchmarking against 13 classical methods. • Mapping relationship between wind power output and meteorological information at a province level is recognized by integrating GIS and CNN. • A new fusion mechanism for geographic and meteorological information is proposed to achieve lightweight modeling. • Comparative experimental studies are conducted in Yunnan province, China. [ABSTRACT FROM AUTHOR]

In recent years, diverse energy has been integrated into the power system, which constitutes a regional integrated energy system (IES). However, the coupling and complementation of multiple energy sources make load forecasting more difficult. For the time‐sequence and non‐linear characteristics of electric load and the complementarity of different energy in IES, this paper proposed an attention‐based convolutional neural network (CNN) combined with long short‐term memory (LSTM) and bidirectional long short‐term memory (BiLSTM) model for short‐term load forecasting in IES. The historical load, temperature, cooling load, and gas consumption of the past 5 days are used as the input features. CNN integrated with attention block is utilized to extract effective features of the load impact factors. Then the load of the next hour is forecasted by the LSTM combined with BiLSTM layers. Finally, the model is verified by the data from an integrated energy park in North China. The results show that the proposed method has better forecasting performance than CNN‐BiLSTM, CNN‐LSTM, BiLSTM, LSTM, backpropagation neural network (BPNN), random forest regression (RFR), and support vector machine regression (SVR). [ABSTRACT FROM AUTHOR]

BACKGROUND Rice blast fungus is a worldwide disease, and it is one of the most serious rice diseases in the north and south rice fields in China. The initial symptoms of rice blast are not obvious, and the speed of transmission is fast. Manual identification is time‐consuming and laborious. At present, it is a great challenge to realize rapid and accurate early identification of rice blast. RESULTS: In this paper, an identification method based on crop disease spores' diffraction fingerprint texture for rice blast was studied; this method utilizes the light field and texture features of diffraction images. To verify the reliability of the model that we proposed, we selected two methods of manual identification and machine recognition to compare and detect rice blast spores. The experimental results show that the identification of light diffraction characteristics is not only higher than the traditional manual recognition by microscope (increased by more than 0.3%), but also faster after neural network training (increased by more than 90%). The diffraction recognition method used in this study, based on crop disease spores' diffraction fingerprint texture, can be completed in a few seconds, and its test accuracy is 97.18%. CONCLUSION: The proposed method, a rapid rice blast detection and identification method based on crop disease spores' diffraction fingerprint texture, has certain advantages compared with the existing manual identification by microscope. This method can be applied to the recognition of rice blast in agricultural research. © 2020 Society of Chemical Industry [ABSTRACT FROM AUTHOR]

Cotton (Gossypium hirsutum) is one of the most important cash crops in China, The timely and effective removal of weeds in cotton seedling stage is an important measure to ensure high and stable yield of cotton. Nowadays, weed recognition based on machine vision is widely used. The fast and effective recognition of crop and weed in the field under natural illumination is one of the key technologies for the development of intelligent mechanization weeding pattern. In the one hand, cotton and weeds have similar color feature in the field. Feature presentation of the natural property of target is difficult to be obtained by the hand-engineered feature extractor. The spatial consistency of the obtained features is not good, and the real-time performance of recognition system is reduced for the complex feature extraction algorithm. On the other hand, the effect of image preprocessing has important influence on recognition results. In order to solve the main problems in the current research, we explored the way to improve the recognition accuracy, stability and real-time performance, and a recognition method of crop and weed based on Faster R-CNN．In this paper, cotton seedling at 2-5 leaves stages and weeding during the same stage were used as research objects under natural illumination. Weed identification from digital images taken under natural illumination at field level is still challenging in agricultural image processing applications, though a lot of research has been conducted related to this topic. To address this problem, images including cottons and weeds were taken vertically from top to bottom. A method based on Faster R-CNN convolutional neural network was proposed to identify weeds from cotton plants more accurately and quickly. The residual network was used to extract image features, with ReLU as the activation function and Max-pooling as the down-sampling method. In the region of proposal network, feature pyramid network was introduced to generate target candidate frame, and Softmax regression classifier was utilized to optimize the CNN network. The proposed methodology was implemented on 200 digital images taken under natural illumination. The experimental results demonstrated that, the average accuracy of weed identification reached 95.5%, and the average time for individual weed plant identification was 1.51 s, which was reduced to 0.09 s by using GPU. To test the efficiency of the proposed methodology, YOLOv3 method was also carried out on the same training and test datasets. The weed identification results were assessed by mean average precision and average precision. The experimental results showed that better performance was achieved by using our proposed methodology, and better identification accuracy was reached as well. This indicated that the proposed method had a good effect on weed detection under natural illumination, and it will greatly promote the development of precise weed control. [ABSTRACT FROM AUTHOR]

PM2.5 Prediction is a complex task of large-scale spatio-temporal analysis, which not only needs comprehension of static geospatial knowledge and relative features but also needs to analyze the real-time situation. This paper discusses the characteristics of the static graph and the dynamic graph in spatio-temporal series tasks. An Adaptive Scalable Spatio-temporal Graph Convolutional Network(ASGCN) model is proposed to predict PM2.5. To capture and analyze the characteristics of the time series period of PM2.5, a time convolution network based on the strategies of inception and gating is proposed and used as a temporal module. A dynamic graph idea is adopted to distinguish the spatio-temporal similarity of different periods. And an adaptive weighted multilayer graph convolution network is used to process static and dynamic graphs, aiming to analyze the spatial relationship of PM2.5 stations. The convolution network with the inception and gating improves the time-series feature capture ability, and adaptive static and dynamic graphs enhance the spatial relationship analysis ability. The temporal and spatial modules of the model are relatively independent, which benefits obtaining the potential information of datasets to improve the prediction accuracy. At the same time, these modules cooperate to make the model adaptable to various data. We choose a great number of comparative models and design a thorough experimental scheme including single-step prediction, multi-step prediction, hyperparameter experiments, and ablation experiments on two real PM2.5 datasets collected in China. Finally, the model achieves performance close to or better than the current state-of-the-art models selected for comparison in prediction tasks. [ABSTRACT FROM AUTHOR]

Seabed sediment classification has important applications in ocean engineering and submarine pipeline design. This paper proposed the dimension-invariant residual network to improve the seabed sediments classification accuracy. The dimension of each channel of the feature extracted from the dimension-invariant residual block is the same as the original image. Compared with down-sampling, each feature vector has more feature parameters, which can improve the accuracy of seabed sediments classification. Shortcut connections were introduced in the proposed network, which can make the network easier to optimize and achieve good performance by using only a not very deep network. It also has a lower complexity compared to powerful classifiers like Transformer and has a better performance in small-scale seabed sediment classification. And to improve the accuracy of seabed sediment classification without consuming too much time for training the classifier, the network parameters optimization method was proposed in this paper. Recording the classification accuracy and training time of the network with these different parameters. Determining the parameters of the network by the proposed optimization method. The proposed classification model and optimization method were validated by using the side-scan sonar images, acquired from the area around the Shandong Peninsula of China. The layers and convolutional kernel sizes were determined to be 8 and 3, respectively. The classification accuracy of overall, rock, sand, and mud of the dimension-invariant residual network model were 97.83%, 94.04%, 97.80%, and 98.02% respectively. In comparison with state-of-the-art classifiers, the experimental results demonstrated that the proposed classification model achieved better performance, and the network parameters optimization method enables the network to achieve satisfactory classification accuracy while spending less time. The proposed classification model and network parameters optimization method show great potential in seabed sediments classification, which can also be extended to the task of underwater target recognition in future research. • The dimension-invariant residual network identifies seabed sediments more accurately. • The dimension-invariant residual block can retain more features than the traditional. • Network parameter optimization methods can balance the training time and performance. [ABSTRACT FROM AUTHOR]

[Display omitted] • We developed an ensemble mineral prospectivity mapping method based on attention-based convolutional neural network. • The proposed ensemble method improves mineral prospectivity modeling accuracy by learning the key channel feature attentively. • Mineral potential map of W–Sn deposits in Nanling metallogenic belt was produced using the proposed method. Data-driven mineral prospectivity mapping (MPM) based on deep learning methods has become a powerful tool for mineral exploration targeting in the past years. Convolutional neural networks (CNNs) have shown great success in this field because of their powerful ability to capture the complex spatial geo-anomalies related to mineralization. However, the exploration big data applied to MPM mainly relies on the high dimensions of evidence layers (other than spatial dimensions), namely, a large number of channels. This impedes the extraction of key channel features related to mineralization when using traditional CNNs. In this paper, we developed an ensemble MPM method based on CNN and Attention model: the ATT–CNN method. Specifically, a channel attention layer is added after the convolution operation of the CNN to enhance the extraction of key channel features in complex exploration data, thereby improving the feature extraction ability and prediction accuracy of CNN for MPM. A case study of W–Sn mineral prospectivity modeling in the Nanling metallogenic belt in South China was used to verify the proposed method. To alleviate the issue of training sample scarcity, we used data augmentation methods (including sliding window and random zero noise addition) when training CNN models. The results show that the prediction accuracies of the ATT–CNN model (92.949% and 94.872% using sliding window and random zero noise addition, respectively) are higher than those of the traditional CNN (91.667% and 92.308%, respectively). Moreover, the improved areas under the receiver operating characteristic curves (AUC) of ATT–CNN (0.987 and 0.971) compared to those of the CNN (0.970 and 0.964) suggest that the proposed ensemble method improves the geological generalization of CNN. The high agreement with known deposits suggests that the areas targeted in this study can guide future mineral exploration of the W–Sn mineralization in the Nanling range. [ABSTRACT FROM AUTHOR]

• A novel rockburst prediction framework is raised based on microseismic monitoring and machine learning. • An integrated CNN-Adam-BO algorithm is proposed to recognize the microseismic sequence. • A field rockburst risk warning system is demonstrated. • Compared to BPNN, SVM, and two benchmark models, the CNN-Adam-BO model is the most superior. Rockburst is a dynamic geological disaster common during underground excavation, which significantly threatens the safety of personnel, equipment, and property. This paper proposes an integrated CNN-Adam-BO algorithm based on microseismic monitoring data to achieve real-time prediction of rockburst intensity. A total of 400 rockburst cases were collected from a water conveyance tunnel in northwestern China, including 100 no rockbursts, 252 low rockbursts, and 48 moderate rockbursts. During modeling, adaptive moment estimation optimization and Bayes optimization were separately used to tune the CNN network hyper-parameters (i.e., convolution kernels and bias) as well as dropout probability, and a parametric analysis was carried out to optimize the CNN structure and training parameters (i.e., the CNN architecture, the number of kernels in the convolutional layer, the number of neurons in the full connection layer, and the number of training epochs). To assess model performance, 2 global metrics (i.e., accuracy and Kappa) and 3 within-class metrics (i.e., precision, recall, and F 1 -score) were introduced. Additionally, a comparison analysis with other machine learning models, such as a back-propagation neural network, a support vector machine, and two CNN-based benchmark models, was conducted in regard to prediction ability and running speed. Results show that the CNN-Adam-BO model is the most superior, which achieved an accuracy of 91.67%, Kappa of 0.8392, F 1 -score of 0.8750, 0.9474, and 0.8333 (on no/low/moderate rockburst), and running time of 190.12 s. Taking 2 extra rockburst cases as examples, a complete workflow for how to apply the constructed model to construct a field rockburst risk warning system was illustrated. Further, the effect of data quality, especially data imbalance, on model performance was discussed, indicating that the synthetic minority oversampling technique is an effective way to eliminate data imbalance. Finally, superiority of the dimension enhancement method of the microseismic sequence described in this paper was verified via comparison with the continuous wavelet transform and cross wavelet transform. [ABSTRACT FROM AUTHOR]

• A control framework based on deep reinforcement learning is designed to realize the real-time power system emergency generator tripping control. • Considering the generator tripping actions as continuous actions, a deep reinforcement learning algorithm, deep deterministic policy gradient, is employed in the framework. • Convolutional neural network is adopted in the deep deterministic policy gradient algorithm to improve the training speed and the generalization ability of the model. In case of faults or severe disturbances, the power system will enter an emergency operation state. After the system instability is detected, oscillation and blackout will occur in the system if effective control measures are not taken in time. Generator tripping control (GTC) is the most effective emergency control measure. In view of the mismatch between the traditional GTC algorithm and the transient stability assessment method based on machine learning, a new real-time GTC method is needed. In this paper, a three-part control framework is designed for the GTC problem. The control agent is endowed with decision-making ability by interacting with the simulation environment in the offline pre-learning part. Then the trained agent is transplanted to the online application which can help system operators make decisions. Meanwhile, the agent is updated with real data to be better adapted to the actual system in the online learning part. A deep reinforcement learning algorithm, deep deterministic policy gradient (DDPG) is employed to train the control agent in this framework. A modified DDPG algorithm and the corresponding reward function are designed for the GTC problem. Convolution neural network (CNN) is added to the DDPG network, by which the training time of the agent is shortened and the generalization ability of the algorithm is improved. Trained with simulation data and real system experience, the control agent can determine control strategies timely according to the system operating conditions. Simulation results on the IEEE-39 bus system and the realistic regional power system of Eastern China show the effectiveness, generalizability, and timeliness of the decision algorithm. [ABSTRACT FROM AUTHOR]

Landslides pose a constant threat to the lives and property of mountain people and may also cause geomorphological destruction such as soil and water loss, vegetation destruction, and land cover change. Landslide susceptibility assessment (LSA) is a key component of landslide risk evaluation. There are many related studies, but few analyses and comparisons of models for optimization. This paper aims to introduce the Tree-structured Parzen Estimator (TPE) algorithm for hyperparameter optimization of three typical neural network models for LSA in Shuicheng County, China, as an example, and to compare the differences of predictive ability among the models in order to achieve higher application performance. First, 17 influencing factors of landslide multiple data sources were selected for spatial prediction, hybrid ensemble oversampling and undersampling techniques were used to address the imbalanced sample and small sample size problem, and the samples were randomly divided into a training set and validation set. Second, deep neural network (DNN), recurrent neural network (RNN), and convolutional neural network (CNN) models were adopted to predict the regional landslides susceptibility, and the TPE algorithm was used to optimize the hyperparameters respectively to improve the assessment capacity. Finally, to compare the differences and optimization effects of these models, several objective measures were applied for validation. The results show that the high-susceptibility regions mostly distributed in bands along fault zones, where the lithology is mostly claystone, sandstone, and basalt. The DNN, RNN, and CNN models all perform well in LSA, especially the RNN model. The TPE optimization significantly improves the accuracy of the DNN and CNN (3.92% and 1.52%, respectively), but does not improve the performance of the RNN. In summary, our proposed RNN model and TPE-optimized DNN and CNN model have robust predictive capability for landslide susceptibility in the study area and can also be applied to other areas containing similar geological conditions. [ABSTRACT FROM AUTHOR]

Heavy rain associated with landfalling typhoons often leads to disasters in South China, which can be reduced by improving the accuracy of radar quantitative precipitation estimation (QPE). At present, raindrop size distribution (DSD)-based nonlinear fitting (QPEDSD) and traditional neural networks are the main radar QPE algorithms. The former is not sufficient to represent the spatiotemporal variability of DSDs through the generalized Z–R or polarimetric radar rainfall relations that are established using statistical methods since such parametric methods do not consider the spatial distribution of radar observables, and the latter is limited by the number of network layers and availability of data for training the model. In this paper, we propose an alternative approach to dual-polarization radar QPE based on deep learning (QPENet). Three datasets of "dual-polarization radar observations—surface rainfall (DPO—SR)" were constructed using radar observations and corresponding measurements from automatic weather stations (AWS) and used for QPENetV1, QPENetV2, and QPENetV3. In particular, 13 × 13, 25 × 25, and 41 × 41 radar range bins surrounding each AWS location were used in constructing the datasets for QPENetV1, QPENetV2, and QPENetV3, respectively. For training the QPENet models, the radar data and AWS measurements from eleven landfalling typhoons in South China during 2017–2019 were used. For demonstration, an independent typhoon event was randomly selected (i.e., Merbok) to implement the three trained models to produce rainfall estimates. The evaluation results and comparison with traditional QPEDSD algorithms show that the QPENet model has a better performance than the traditional parametric relations. Only when the hourly rainfall intensity is less than 5 mm (R < 5 mm·h−1), the QPEDSD model shows a comparable performance to QPENet. Comparing the three versions of the QPENet model, QPENetV2 has the best overall performance. Only when the hourly rainfall intensity is less than 5 mm (R < 5 mm·h−1), QPENetV3 performs the best. [ABSTRACT FROM AUTHOR]

Power dispatching systems currently receive massive, complicated, and irregular monitoring alarms during their operation, which prevents the controllers from making accurate judgments on the alarm events that occur within a short period of time. In view of the current situation with the low efficiency of monitoring alarm information, this paper proposes a method based on natural language processing (NLP) and a hybrid model that combines long short-term memory (LSTM) and convolutional neural network (CNN) for the identification of grid monitoring alarm events. Firstly, the characteristics of the alarm information text were analyzed and induced and then preprocessed. Then, the monitoring alarm information was vectorized based on the Word2vec model. Finally, a monitoring alarm event identification model based on a combination of LSTM and CNN was established for the characteristics of the alarm information. The feasibility and effectiveness of the method in this paper were verified by comparison with multiple identification models. [ABSTRACT FROM AUTHOR]

The well testing technique is an important tool in estimating well and reservoir characteristics, such as permeability, skin factor and so on. For a long time, researchers have been searching for automatic well testing interpretation tools, but the results are disappointing. This paper proposes using convolutional neural network (CNN) as an automatic well test interpretation approach for infinite acting reservoirs. The CNN takes pressure change and pressure derivative data of the log-log plot for inputs. The wellbore storage coefficient, skin factor and reservoir permeability are redefined into a dimensionless group C D e 2 S as the output of the CNN. In this method, the best trained CNN structure is obtained by minimizing mean square error (MSE) and mean relative error (MRE). This new method is tested for its effectiveness and accuracy in Daqing oil field, China. It demonstrates that, for wells in infinite reservoir, CNN could be an effective automatic well test interpretation technique. CNN also shows the potential for more complicated scenarios. • A well test interpretation method based on convolutional neural network (CNN) is proposed. • The log-log plot is taken as the input of the CNN, and the log (C D e 2 S) is used as the CNN's output. • The trained CNN can interpret pressure data measured in shut-in well effectively and accurately. [ABSTRACT FROM AUTHOR]

Diagnosis of benign–malignant nodules in the lung on Computed Tomography (CT) images is critical for determining tumor level and reducing patient mortality. Deep learning-based diagnosis of nodules in lung CT images, however, is time-consuming and less accurate due to redundant structure and the lack of adequate training data. In this paper, a novel diagnosis method based on Deep Transfer Convolutional Neural Network (DTCNN) and Extreme Learning Machine (ELM) is explored, which merges the synergy of two algorithms to deal with benign–malignant nodules classification. An optimal DTCNN is first adopted to extract high-level features of lung nodules, which has been trained with the ImageNet dataset beforehand. After that, an ELM classifier is further developed to classify benign and malignant lung nodules. Two datasets, including the Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI) public dataset and a private dataset from the First Affiliated Hospital of Guangzhou Medical University in China (FAH-GMU), have been conducted to verify the efficiency and effectiveness of the proposed approach. For LIDC-IDRI dataset, the experimental results show that our novel DTCNN-ELM model achieved the performance with an accuracy of 94.57%, a sensitivity of 93.69%, a specificity of 95.15%, an area under the receiver operator curve (AUC) of 94.94%, and testing time per nodule of 0.5 ms, which has the most reliable results compared with current state-of-the-art methods. Codes are available 1 1 https://github.com/hxf1228/dtcnn_elm_lnd.. • Deep Transfer Convolutional Neural Networks (DTCNN) is used for nodule diagnosis. • An optimal pre-trained DTCNN is adopted to capture richer features of lung nodules. • An Extreme Learning Machine (ELM) model is used to ease the computational burden. • Two cases study shows the effectiveness and efficiency of the proposed approach. [ABSTRACT FROM AUTHOR]

Accurate degradation tendency prediction (DTP) is vital for the secure operation of a pumped storage unit (PSU). However, the existing techniques and methodologies for DTP still face challenges, such as a lack of appropriate degradation indicators, insufficient accuracy, and poor capability to track the data fluctuation. In this paper, a hybrid model is proposed for the degradation tendency prediction of a PSU, which combines the integrated degradation index (IDI) construction and convolutional neural network-long short-term memory (CNN-LSTM). Firstly, the health model of a PSU is constructed with Gaussian process regression (GPR) and the condition parameters of active power, working head, and guide vane opening. Subsequently, for comprehensively quantifying the degradation level of PSU, an IDI is developed using entropy weight (EW) theory. Finally, combining the local feature extraction of the CNN with the time series representation of LSTM, the CNN-LSTM model is constructed to realize DTP. To validate the effectiveness of the proposed model, the monitoring data collected from a PSU in China is taken as case studies. The root mean square error (RMSE), mean absolute error (MAE) and mean absolute percentage error (MAPE) obtained by the proposed model are 1.1588, 0.8994, 0.0918, and 0.9713, which can meet the engineering application requirements. The experimental results show that the proposed model outperforms other comparison models. [ABSTRACT FROM AUTHOR]

How to acquire landslide disaster information quickly and accurately has become the focus and difficulty of disaster prevention and relief by remote sensing. Landslide disasters are generally featured by sudden occurrence, proposing high demand for emergency data acquisition. The low-altitude Unmanned Aerial Vehicle (UAV) remote sensing technology is widely applied to acquire landslide disaster data, due to its convenience, high efficiency, and ability to fly at low altitude under cloud. However, the spectrum information of UAV images is generally deficient and manual interpretation is difficult for meeting the need of quick acquisition of emergency data. Based on this, UAV images of high-occurrence areas of landslide disaster in Wenchuan County and Baoxing County in Sichuan Province, China were selected for research in the paper. Firstly, the acquired UAV images were pre-processed to generate orthoimages. Subsequently, multi-resolution segmentation was carried out to obtain image objects, and the barycenter of each object was calculated to generate a landslide sample database (including positive and negative samples) for deep learning. Next, four landslide feature models of deep learning and transfer learning, namely Histograms of Oriented Gradients (HOG), Bag of Visual Word (BOVW), Convolutional Neural Network (CNN), and Transfer Learning (TL) were compared, and it was found that the TL model possesses the best feature extraction effect, so a landslide extraction method based on the TL model and object-oriented image analysis (TLOEL) was proposed; finally, the TLOEL method was compared with the object-oriented nearest neighbor classification (NNC) method. The research results show that the accuracy of the TLOEL method is higher than the NNC method, which can not only achieve the edge extraction of large landslides, but also detect and extract middle and small landslides accurately that are scatteredly distributed. [ABSTRACT FROM AUTHOR]