Search

The identification of damage to concrete bridge surfaces is of great significance to maintaining the durability and reliability of bridges. However, it is difficult to identify small areas of damage, especially in high-resolution images. To conduct damage identification research in a targeted manner, this paper proposes a method of multidamage identification in high-resolution images based on You Only Look Once version 5 (YOLOv5). In this paper, a data set labeled with four types of damage (crack, spallation, hole, and rebar) is used for training, validation, and testing. To ensure that the network adequately learns the damage features, the high-resolution images are cropped into subimages via the autoadaptive window cropping method (AWCM) proposed in this paper. The cropping method can crop images according to the label information and protect the damage features from destruction during the cropping process. To avoid overfitting, it is necessary to balance the volume of different types of damage. After balancing, the count for each category is as follows: crack (4,980), spallation (5,225), hole (5,211) and rebar (5,020). The balanced data set can be used to train the deep learning network and construct the multidamage identification model. After identification, the subimages, and the prediction boxes of damages in them are restored to their original high-resolution images. The results show that the mean average precision (mAP) of all classes is 94.2%, and the values for cracks, spallation, holes, and rebars are 86.9%, 98.1%, 92.3%, and 99.4%, respectively, which indicates that the proposed method outperforms the other three methods (inputting original images directly, the sliding window cropping method, and the random centroid cropping method). [ABSTRACT FROM AUTHOR]

In the field of medical image analysis, object detection plays a crucial role by providing interpretable diagnostic information to healthcare professionals. Although current object detection models have achieved remarkable success in conventional images, their performance in detecting abnormalities in medical images has not been as satisfactory. This is primarily due to the complexity of anatomical structures in medical images, and the fact that some lesions may have subtle features, particularly in the case of early‐stage, small‐scale abnormalities. To address this challenge, we introduce SOCR‐YOLO, a novel lesion detection model with online convolutional reparameterization based on channel shuffling. First, it employs the SOCR (Shuffled Channel with Online Convolutional Re‐parameterization) module to establish a connection between feature concatenation and computational efficiency, aiming to extract more comprehensive information while reducing time consumption. Second, it incorporates the Bi‐FPN structure to achieve multiscale feature fusion. Lastly, the loss function has been optimized to improve the model training process. We evaluated two datasets, chest x‐ray (Vindr‐CXR) and brain tumor (Br35H), provided by the Kaggle competition. Experimental results show that the proposed method has outperformed several state‐of‐the‐art models, including YOLOv8, YOLO‐NAS, and RT‐DETR, in both speed and accuracy. Notably, in the context of chest x‐ray anomaly detection, SOCR‐YOLO exhibits a 1.8% enhancement in accuracy over YOLOv8 while simultaneously reducing floating‐point operations by 26.3%. Additionally, a similar 1.8% improvement in accuracy is observed in the detection of brain tumors. The results indicate the superior ability of our model to detect multiscale variations and small lesions. [ABSTRACT FROM AUTHOR]

High-resolution remote sensing image fusion technology can combine multiple remote sensing images to display images more richly and accurately, and make them present more comprehensive and accurate image features. Because this technology has a high utilization rate of image processing, it has research significance. The objective of this study is to improve the accuracy and reliability of infrared image processing technology in motion recognition. In order to achieve this goal, this paper focuses on the research of high resolution image processing technology to improve the quality and clarity of infrared images. Image enhancement algorithm is used to reduce the noise and blur effect in infrared images, image recovery technology is used to improve the resolution and detail representation ability of infrared images, and finally image segmentation and feature extraction algorithm are used to extract the motion features in infrared images. [ABSTRACT FROM AUTHOR]

We introduce a highly efficient 48-channel ultrasound beamforming system ideal for ultrasound endoscopy applications. The system includes a transmitter and a receiver that allows for low-area, high-resolution imaging acquisition. The transmitter uses a charge redistribution HV (high-voltage) scheme to generate three-level pulses that actuate the transducer, implemented with the standard CMOS process for optimal cost and power savings. Meanwhile, the receiver features a sub-array structure and a delay generator that reduces the area usage. To achieve high-resolution ultrasound imaging acquisition with low computational power, we developed a Shift Coherence Factor (SCF) algorithm that is hardware-friendly. This approach delivers a lateral resolution of over 20% better than that of the conventional delay and sum (DAS) algorithm, with a contrast ratio of over 30 dB. The system was implemented in a 180 nm standard CMOS process with an area of 24.98 mm2, power consumption of 8.23 mW per channel, achieving a delay resolution of 8.33 ns, and a low-area implementation of 0.52 mm2 per channel. The system offers high-quality imaging acquisition with minimal additional area and power consumption, which has great potential for 3D imaging or catheterized ultrasound systems. [ABSTRACT FROM AUTHOR]

Histopathology images are visual representations of tissue samples that have been processed and examined under a microscope in order to establish diagnoses for various disorders. These images are categorized by deep transfer learning due to the absence of big annotated datasets. There are some classifiers such as softmax and Support Vector Machine (SVM) used to perform multiple and binary classification respectively. Feature reduction for high dimensional images, is an emerging technique which can meet two basic criteria’s of classification i.e. it deals with over-fitting issue and it can also incredibly increase the classification accuracy. As disease diagnosis requires accurate histopathological image classification, so the proposed Dual Featured Auto-encoder (DFAE) based transfer learning is introduced with Triple Layered Convolutional Architecture. The Histological CIMA dataset is used after pre-processing by PHAT, a mathematical and computational framework to get spatial features as well as spectral features. In order to achieve the two objectives, the proposed integrated methodology uses reduced informative features from DFAE and fed them to Triple Layered Convolutional Architecture (TLCA). The conventional Convolutional Neural Network (CNN), ResNet50, Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) are also tested against reduced dimensional image data but we found moderate or even low accuracies i.e. 25% for DFAE-ResNet50, 66% for DFAE-LSTM, 33% for DFAE-GRU and 67% for DFAE-CNN. While the accuracy of our proposed architecture Dual Featured Auto-encoder with TLCA (DFAE-TLCA) is better i.e. 96.07%. The proposed methodology has the potential to revolutionize the medical research.

The rapid pace of urbanization and increasing demands for urban functionalities have led to diversification and complexity in the types of urban surface elements. The conventional approach of relying solely on remote sensing imagery for urban surface element extraction faces emerging challenges. Data-driven techniques, including deep learning and machine learning, necessitate a substantial number of annotated samples as prerequisites. In response, our study proposes a knowledge-driven approach that integrates multisource data with ontology to achieve precise urban surface element extraction. Within this framework, components from the EIONET Action Group on Land Monitoring in Europe matrix serve as ontology primitives, forming a shared vocabulary. The semantics of surface elements are deconstructed using these primitives, enabling the creation of specific descriptions for various types of urban surface elements by combining these primitives. Our approach integrates multitemporal high-resolution remote sensing data, network big data, and other heterogeneous data sources. It segments high-resolution images into individual patches, and for each unit, urban surface element classification is accomplished through semantic rule-based inference. We conducted experiments in two regions with varying levels of urban scene complexity, achieving overall accuracies of 93.03% and 97.35%, respectively. Through this knowledge-driven approach, our proposed method significantly enhances the classification performance of urban surface elements in complex scenes, even in the absence of sample data, thereby presenting a novel approach to urban surface element extraction. [ABSTRACT FROM AUTHOR]

Seismic interpretation aims to extract quantitative and interpretable attributes from a seismic image produced using some migration method to inform characteristics of a subsurface reservoir or target of interest. Current paradigms for computing seismic attributes mostly rely on single-task algorithms. We develop an iterative, multitask machine learning method to learn and infer multiple attributes from a seismic image. This method is composed of two stages: a multitask inference stage and a multimodal, multitask refinement stage. The basic mechanism of this method is that we train a multitask inference neural network to estimate a set of attributes, including a relative geological time volume, a denoised higher-resolution seismic image and multiple fault attributes (including probability, dip and strike), from a low-resolution, noisy seismic image; then we input the inferred attributes to a multitask refinement NN to enhance the raw inference results iteratively. The two multitask neural networks are trained separately based on synthetic seismic images and associated labels generated by geological modelling. Applications of this multitask learning and inference method to synthetic and field seismic images show that our method can improve the structural consistency among output seismic attributes compared with single-task neural networks, leading to more reliable automatic interpretation and subsurface characterization. [ABSTRACT FROM AUTHOR]

In March 1966, Ms 6.8 and 7.2 earthquakes occurred in Xingtai, North China, resulting in widespread soil liquefaction that caused severe infrastructure damage and economic losses. Using Keyhole satellite imagery combined with aerial images and fieldwork records, we interpreted and identified 66,442 liquefaction points and analyzed the coseismic liquefaction distribution characteristics and possible factors that influenced the Xingtai earthquakes. The interpreted coseismic liquefaction was mainly concentrated above the IX-degree zone, accounting for 80% of all liquefaction points. High-density liquefaction zones (point density > 75 pieces/km2) accounted for 22% of the total liquefaction points. Most of the interpreted liquefaction points were located at the region with a peak ground acceleration (PGA) of >0.46 g. The liquefaction area on 22 March was significantly larger than that on 8 March. The region of liquefaction was mainly limited by sandy soil conditions, water system conditions, and seismic geological conditions and distributed in areas with loose fine sand and silt deposits, a high water table (groundwater level increases before both mainshocks corresponding to the liquefaction intensive regions), rivers, and ancient river channels. Liquefaction exhibited a repeating characteristic in the same region. Further understanding of the liquefaction characteristics of Xingtai can provide a reference for the prevention of liquefaction in northern China. [ABSTRACT FROM AUTHOR]

Parcel-level farmland information contains rich spatial distribution and boundary details, which is crucial for digital agriculture and agricultural resource surveys. However, the spatial complexity and heterogeneity of features resulting from high resolution makes it difficult to obtain parcel-level information quickly and accurately. In addition, existing methods do not sufficiently take into account the spatial topological information, particularly for blurred boundaries. Here, we develop a multi-task network model to extract plot-level cropland information. Specifically, the model consists of a cascaded multi-task network with integrated semantic and edge detection, a refinement network with fixed edge local connectivity, and an integrated fusion model. To validate the performance of the model, two typical tests were conducted in Denmark (Europe) and Chongqing (Asia) with high-resolution remote sensing images provided by Sentinel-2 (10 m) and Google Earth (0.53 m) as data sources. The results show that our proposed model outperforms other baseline models and exhibits higher performance. This study is expected to provide important support for the design of new global agricultural information management systems in the future. [ABSTRACT FROM AUTHOR]

ABSTRACTParcel-level farmland information contains rich spatial distribution and boundary details, which is crucial for digital agriculture and agricultural resource surveys. However, the spatial complexity and heterogeneity of features resulting from high resolution makes it difficult to obtain parcel-level information quickly and accurately. In addition, existing methods do not sufficiently take into account the spatial topological information, particularly for blurred boundaries. Here, we develop a multi-task network model to extract plot-level cropland information. Specifically, the model consists of a cascaded multi-task network with integrated semantic and edge detection, a refinement network with fixed edge local connectivity, and an integrated fusion model. To validate the performance of the model, two typical tests were conducted in Denmark (Europe) and Chongqing (Asia) with high-resolution remote sensing images provided by Sentinel-2 (10 m) and Google Earth (0.53 m) as data sources. The results show that our proposed model outperforms other baseline models and exhibits higher performance. This study is expected to provide important support for the design of new global agricultural information management systems in the future.

Hyperspectral satellite imagery has developed rapidly over the last decade because of its high spectral resolution and strong material recognition capability. Nonetheless, the spatial resolution of available hyperspectral imagery is inferior, severely affecting the accuracy of ground object identification. In the paper, we propose an adaptively optimized pulse-coupled neural network (PCNN) model to sharpen the spatial resolution of the hyperspectral imagery to the scale of the multispectral imagery. Firstly, a SAM-CC strategy is designed to assign hyperspectral bands to the multispectral bands. Subsequently, an improved PCNN (IPCNN) is proposed, which considers the differences of the neighboring neurons. Furthermore, the Chameleon Swarm Optimization (CSA) optimization is adopted to generate the optimum fusion parameters for IPCNN. Hence, the injected spatial details are acquired in the irregular regions generated by the IPCNN. Extensive experiments are carried out to validate the superiority of the proposed model, which confirms that our method can realize hyperspectral imagery with high spatial resolution, yielding the best spatial details and spectral information among the state-of-the-art approaches. Several ablation studies further corroborate the efficiency of our method. [ABSTRACT FROM AUTHOR]

Deep learning models have enabled significant performance improvements to remote sensing image processing. Usually, a large number of training samples is required for detection models. In this study, a dynamic simulation training strategy is designed to generate samples in real time during training. The few adversarial examples are not only directly involved in the training but are also used to fit the distribution model of adversarial noise, helping the real-time generated samples to be similar to adversarial examples. The noise of the training samples is randomly generated according to the distribution model, and the random variation of training inputs reduces the overfitting phenomenon. To enhance the detectability of adversarial noise, the input model is no longer a normalized image but a JPEG error image. Experiments show that with the proposed dynamic simulation training strategy, common classification models such as ResNet and DenseNet can effectively detect adversarial examples. [ABSTRACT FROM AUTHOR]

To solve the problem of high-resolution image alignment time overhead, an SIFT-based fast image alignment algorithm is presented. The overlap region of images is computed in detail by phase correlation algorithm to avoid a lot of useless calculations of non-overlapping region. After the distribution of feature points determined in difference of Gaussian through formula derivation, the total number of feature points is limited. The more stable spatially distributed for the feature points is obtained due to the expanded detection range of extreme points and added non-maximum suppression. It is noteworthy that the range of the descriptor is calculated by the method of down-sampling. And the circular descriptor is constructed with only 56-dimensional in the feature point descriptor generation stage, which makes the time of the descriptor generation and feature point matching shorter. This indicates that the total descriptor calculation is faster in lower dimensions by the new algorithm. In addition, experimental results show that the average time (9.60s, 13.46s, and 15.81s) of the proposed algorithm is 0.86%, 0.43%, and 0.10% of the SIFT algorithm, respectively. The overall speed is 2–3 orders of magnitude faster than the SIFT algorithm, which indicates that the new algorithm can solve the problem of high-resolution image alignment time overhead. The new algorithm provides a good stitching quality and shows an excellent performance for high-resolution image compared with several existing image stitching algorithms at the current. It indicates that the algorithm has potential application value in real-time image stitching.

High-resolution deep-learning-based remote-sensing imagery analysis has been widely used in land-use and crop-classification mapping. However, the influence of composite feature bands, including complex feature indices arising from different sensors on the backbone, patch size, and predictions in transferable deep models require further testing. The experiments were conducted in six sites in Henan province from 2019 to 2021. This study sought to enable the transfer of classification models across regions and years for Sentinel-2A (10-m resolution) and Gaofen PMS (2-m resolution) imagery. With feature selection and up-sampling of small samples, the performance of UNet++ architecture on five backbones and four patch sizes was examined. Joint loss, mean Intersection over Union (mIoU), and epoch time were analyzed, and the optimal backbone and patch size for both sensors were Timm-RegNetY-320 and 256 × 256, respectively. The overall accuracy and F1 scores of the Sentinel-2A predictions ranged from 96.86% to 97.72% and 71.29% to 80.75%, respectively, compared to 75.34%–97.72% and 54.89%–73.25% for the Gaofen predictions. The accuracies of each site indicated that patch size exerted a greater influence on model performance than the backbone. The feature-selection-based predictions with UNet++ architecture and up-sampling of minor classes demonstrated the capabilities of deep-learning generalization for classifying complex ground objects, offering improved performance compared to the UNet, Deeplab V3+, Random Forest, and Object-Oriented Classification models. In addition to the overall accuracy, confusion matrices, precision, recall, and F1 scores should be evaluated for minor land-cover types. This study contributes to large-scale, dynamic, and near-real-time land-use and crop mapping by integrating deep learning and multi-source remote-sensing imagery

Copyright of Journal of Remote Sensing is the property of Editorial Office of Journal of Remote Sensing & Science Publishing Co. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In March 1966, Ms 6.8 and 7.2 earthquakes occurred in Xingtai, North China, resulting in widespread soil liquefaction that caused severe infrastructure damage and economic losses. Using Keyhole satellite imagery combined with aerial images and fieldwork records, we interpreted and identified 66,442 liquefaction points and analyzed the coseismic liquefaction distribution characteristics and possible factors that influenced the Xingtai earthquakes. The interpreted coseismic liquefaction was mainly concentrated above the IX-degree zone, accounting for 80% of all liquefaction points. High-density liquefaction zones (point density > 75 pieces/km2) accounted for 22% of the total liquefaction points. Most of the interpreted liquefaction points were located at the region with a peak ground acceleration (PGA) of >0.46 g. The liquefaction area on 22 March was significantly larger than that on 8 March. The region of liquefaction was mainly limited by sandy soil conditions, water system conditions, and seismic geological conditions and distributed in areas with loose fine sand and silt deposits, a high water table (groundwater level increases before both mainshocks corresponding to the liquefaction intensive regions), rivers, and ancient river channels. Liquefaction exhibited a repeating characteristic in the same region. Further understanding of the liquefaction characteristics of Xingtai can provide a reference for the prevention of liquefaction in northern China.

The rapid pace of urbanization and increasing demands for urban functionalities have led to diversification and complexity in the types of urban surface elements. The conventional approach of relying solely on remote sensing imagery for urban surface element extraction faces emerging challenges. Data-driven techniques, including deep learning and machine learning, necessitate a substantial number of annotated samples as prerequisites. In response, our study proposes a knowledge-driven approach that integrates multisource data with ontology to achieve precise urban surface element extraction. Within this framework, components from the EIONET Action Group on Land Monitoring in Europe matrix serve as ontology primitives, forming a shared vocabulary. The semantics of surface elements are deconstructed using these primitives, enabling the creation of specific descriptions for various types of urban surface elements by combining these primitives. Our approach integrates multitemporal high-resolution remote sensing data, network big data, and other heterogeneous data sources. It segments high-resolution images into individual patches, and for each unit, urban surface element classification is accomplished through semantic rule-based inference. We conducted experiments in two regions with varying levels of urban scene complexity, achieving overall accuracies of 93.03% and 97.35%, respectively. Through this knowledge-driven approach, our proposed method significantly enhances the classification performance of urban surface elements in complex scenes, even in the absence of sample data, thereby presenting a novel approach to urban surface element extraction.

In this study, high resolution (HR) images, GIS and deep convolution neural network (DCNN) are used for assessment of traffic congestion. A DCNN architecture comprises one convolution layer, two pooling layers and a five-layer fully connected neural network (FCNN) evaluated for identifying vehicles in a movable window in HR images. A simple mathematical method is followed for changing the scale and orientation of the movable window to optimally mask and measure the area of vehicles. A formula is appraised to compute the traffic density using the estimated vehicles and road areas. A threshold is used to estimate traffic congestion from the measured traffic density. The method is validated by applying on World View-2 pan-sharpened multispectral images having spatial resolution 0.46 m. In comparison to CNN and ResNet-18, the proposed approach achieves a quite promising accuracy (99%) and needs less training and processing time for measuring traffic congestion in HR images. [ABSTRACT FROM AUTHOR]

Hyperspectral satellite imagery has developed rapidly over the last decade because of its high spectral resolution and strong material recognition capability. Nonetheless, the spatial resolution of available hyperspectral imagery is inferior, severely affecting the accuracy of ground object identification. In the paper, we propose an adaptively optimized pulse-coupled neural network (PCNN) model to sharpen the spatial resolution of the hyperspectral imagery to the scale of the multispectral imagery. Firstly, a SAM-CC strategy is designed to assign hyperspectral bands to the multispectral bands. Subsequently, an improved PCNN (IPCNN) is proposed, which considers the differences of the neighboring neurons. Furthermore, the Chameleon Swarm Optimization (CSA) optimization is adopted to generate the optimum fusion parameters for IPCNN. Hence, the injected spatial details are acquired in the irregular regions generated by the IPCNN. Extensive experiments are carried out to validate the superiority of the proposed model, which confirms that our method can realize hyperspectral imagery with high spatial resolution, yielding the best spatial details and spectral information among the state-of-the-art approaches. Several ablation studies further corroborate the efficiency of our method.

Building extraction plays an important role in high-resolution remote sensing image processing, which can be used as the basis for urban planning and demographic analysis. In recent years, many powerful general semantic segmentation models have emerged, but these models often perform poorly when transferred to remote sensing images because of the characteristics of remote sensing images. To this end, we propose a new deep learning network called Selective Nonlocal ResUNeXt++ (SNLRUX++) for building extraction. First, the cascaded multiscale feature fusion is proposed to transform the high-performance image classification network ResNeXt into the segmentation network ResUNeXt++. Second, selective nonlocal operation is designed to establish long-range dependencies while avoiding introducing excessive noise and computational effort. Finally, multiscale prediction is applied as deep supervision to accelerate training and convergence, and improves prediction performance of objects at different scales. The experimental results on two different remote sensing image datasets show the effectiveness and generalization ability of the proposed method.

Landslides have caused tremendous damage to human lives and property safety. However, the complex environment of mountain landslides and the vegetation coverage around landslides make it difficult to identify landslides quickly and efficiently using high-resolution images. To address this challenge, this article presents a feature-based constraint deep U-Net (FCDU-Net) method to detect rainfall-induced mountainous landslides. Usually, the vegetation in the landslide area is severely damaged, and the vegetation coverage can indirectly reflect the spatial extent of the landslide. Meanwhile, the texture features of high-resolution images can characterize the surface environment of landslide hazards to a certain extent. We first introduce auxiliary features of normalized difference vegetation index and gray-level co-occurrence matrix into the proposed method to further improve the detection performance. Then, to minimize the information redundancy of these features and the image, we combine Relief-F and Deep U-Net to screen the optimal features to effectively identify accurate and detailed landslide boundaries. Compared with traditional semantic segmentation methods, the FCDU-Net method can capture fine-grained details in high-resolution images and produce more accurate segmentation results. We conducted experiments by applying the proposed method and other most popular semantic segmentation methods to a high-resolution RapidEye image in Rio de Janeiro, Brazil. The results demonstrate that the FCDU-Net method can achieve better landslide detection results than the other semantic segmentation methods, and the evaluation measures of Precision, F1 score, and mean Intersection-over-Union are as high as 88.87%, 81.17%, and 83.19%, respectively. Furthermore, we quantitatively analyze the effect of the convolution input window size on the performance of FCDU-Net in detecting landslides. We believe that FCDU-Net can serve as a reliable tool for fast and accurate regional landslide hazard surveys.

In recent years, the automatic ground object extraction from remote sensing images has been dramatically advanced by the fully convolutional networks (FCNs). It is an effective method to fuse high-resolution images and light detection and ranging (LiDAR) data in FCNs to improve the extraction accuracy and the robustness. However, the existing FCN-based fusion networks still face challenges in efficiency and accuracy. Methods: We propose a knowledge-aware and multi-scale fusion network (KMF‐ Net) for robust and accurate ground object extraction. The proposed network incorporates a knowledge aware module in the network encoder for better exploiting remote sensing knowledge between pixels. A series-parallel hybrid convolution module is developed to enhance multi-scale representative features from ground objects. Moreover, the network decoder uses a gradual bilinear interpolation strategy to obtain fine grained extraction results. Results: We evaluate KMFNet in the LuoJiaNET with four current mainstream ground object extraction methods (GRRNet, V-FuseNet, DLR and Res-U-Net) on ISPRS 2D semantic segmentation dataset. The comparative evaluation results show that KMFNet can obtain the best overall ac‐ curacy. Compared with the other four methods, KMFNet achieves a better effect by improving the overall accuracy by 3.20% and 2.82% on average in ISPRS-Vaihingen dataset and ISPRS-Potsdam dataset, re‐ spectively. Conclusions: KMFNet achieves the best extraction results by capturing the intrinsic pixel relationships and strengths the multi-scale representative and detailed features of ground objects. It shows great potential in high-precision remote sensing mapping applications. [ABSTRACT FROM AUTHOR]

Human–computer communication using hand gestures has always been difficult. More than half a century ago, people used differentways of interactionwith computers from the early mediums such as perforated game cards. Nowadays, if a richer lexicon of gestures is given, people can communicate more effectively with computers. Machine learning is now used to recognize and classify hand gestures in amore preciseway. In order to increase the communication between computers and humans, we proposed a technique, which uses a wearable low-cost device to generate the electrical impedance tomography (EIT) images to recover the inner impedance structure of a user’s wrist. This is done by measuring the transverse impedance between all the 16 pairs of electrodesofwrist band that lie on the skin of the user hand. The proposedtechnique is enough to integrate the technology into the prototypewrist band tomonitor and classify gestures in real time. We have conducted a study of 16 gestures with a focus on gross hand and pinch finger gestures. The results evaluation shows that the gross hand gestures achieved 90% accuracy inwrist position, while pinch gestures achieved 93% accuracy. [ABSTRACT FROM AUTHOR]

To solve the problem of low accuracy and slow speed of drone detection in high-resolution images with fixed cameras, we propose a detection method combining background difference and lightweight network SAG-YOLOv5s. First, background difference is used to extract potential drone targets in high-resolution images, eliminating most of the background to reduce computational overhead. Secondly, the Ghost module and SimAM attention mechanism are introduced on the basis of YOLOv5s to reduce the total number of model parameters and improve feature extraction, and α-DIoU loss is used to replace the original DIoU loss to improve the accuracy of bounding box regression. Finally, to verify the effectiveness of our method, a high-resolution drone dataset is made based on the public data set. Experimental results show that the detection accuracy of the proposed method reaches 97.6%, 24.3 percentage points higher than that of YOLOv5s, and the detection speed in 4K video reaches 13.2 FPS, which meets the actual demand and is significantly better than similar algorithms. It achieves a good balance between detection accuracy and detection speed and provides a method benchmark for high-resolution drone detection under a fixed camera. [ABSTRACT FROM AUTHOR]

Objectives: The scale of buildings and their distribution is key indicators to measure the economic and social development of a region. Therefore, it is significant to study the extraction of buildings based on remote sensing images. Existing neural network methods still have shortcomings in the completeness of building extraction and the accuracy of building edges. To solve the above problems, this paper proposes a multi-level feature fusion network (MFFNet) based on high-resolution images. Methods: Firstly, we use edge detection operators to improve the ability of the network to recognize the boundaries of buildings. Secondly, we use a multi-path convolution fusion module to extract building features from multiple dimensions, and introduce a large receptive field convolution module to break through feature extraction. The process is limited by the size of the receptive field. After fusing the extracted features, the convolutional attention module is used to compress them, and the global features are further mined by pyramid pooling, so as to achieve high-precision extraction of buildings. Results: The current mainstream UNet, pyramid scene parsing network (PSPNet),multi attending path neural network (MAPNet) and multiscale-feature fusion deep neural networks with dilated convolution (MDNNet) are used as the comparison methods, and we use Wuhan University Aerial Image Dataset, Satellite Dataset II (East Asia) and Inria Aerial Image Dataset as experimental data for testing. Compared with the other four methods, MFFNet improves intersection over union, precision, recall, F1-score and mean average precision by 1.53%, 2.65%, 2.41%, 3.32% and 1.19% on average, achieves a better effect. Conclusions: MFFNet not only accurately captures the detail features of buildings, but also strengthens the extraction and utilization of global features. It has better ex‐ traction effect on large buildings and buildings in complex environment. [ABSTRACT FROM AUTHOR]

Panoptic segmentation is a recent and powerful task that tackles individual object recognition (“things”) and multiple backgrounds (“stuff”) simultaneously. Remote sensing studies with panoptic segmentation are still restricted and recent, with great application perspectives. In this sense, we propose the first multispectral panoptic segmentation study, considering the “thing” and “stuff” classes in the beach scenario and evaluating different sets of spectral bands. Our methodology included developing a dataset with 3800 (3200 for training, 300 for validation, and 300 for testing) with 128 × 128 spatial dimensions and eight spectral bands considering fourteen classes (6 “thing” and 8 “stuff” classes). We used WorldView-3 images from Praia do Futuro, Fortaleza, and pan-sharpening to improve spatial resolution. Five different spectral band configurations were considered: (1) all eight bands, (2) RGB+NIR1+NIR2, (3) RGB+NIR1, (4) RGB+NIR2, and (5) only RGB. The model training used the Panoptic-FPN architecture with the same hyperparameter settings considering three backbones (ResNeXt-101, ResNet-101, and ResNet-50). The best result considered the ResNeXt-101 with all spectral bands. However, the results from the first four configurations were very similar, and the RGB alone was the only configuration with significantly lower results. We also evaluated 15 semantic segmentation models for a benchmark comparison for the Beach Dataset. We show in visual results that even though the semantic models may be precise, they fail at identifying unique targets, especially in crowded locations such as the beach. The panoptic segmentation allowed a necessary detailing and counting of tourist infrastructures and mapping of other background features, establishing an essential tool for inspecting beach areas.

Road network extraction plays a critical role in data updating, urban development, and decision support. To improve the efficiency of labeling road datasets and addressing the problems of traditional methods of manually extracting road networks from high-resolution images, such as their slow speed and heavy workload, this paper proposes a semi-automatic method of road network extraction from high-resolution remote-sensing images. The proposed method needs only a few points to extract a single road in the image. After the roads are extracted one by one, the road network is generated according to the width of each road and the spatial relationships among the roads. For this purpose, we use regional growth, morphology, vector tracking, vector simplification, endpoint modification, road connections, and intersection connections to generate road networks. Experiments on four images with different terrains and different resolutions show that this method has high extraction accuracy under different image conditions. The comparisons with the semi-automatic GVF-snake method based on regional growth also showed its advantages and potentiality. The proposed method is a novel form of semi-automatic road network extraction, and it significantly increases the efficiency of road network extraction. [ABSTRACT FROM AUTHOR]

Transformer-based detectors have recently achieved remarkable success in object detection, revolutionizing the field with their efficiency and accuracy. However, applying these models to high-resolution images presents significant challenges due to the increased computational demands and complexity of processing dense, high-resolution data. In this paper, we introduce a novel model specifically designed for object detection from high-resolution imagery. This model incorporates a multi-layer object-focus network along with a transformer encoder-decoder structure. Specifically, the model employs a dual-head strategy in the object-focus network to balance the detailed analysis of small objects with computational efficiency. This is achieved by leveraging data sparsity to reduce unnecessary computations in massive background regions. Additionally, to improve detection performance for small objects, we propose a method to effectively apply the transformer encoder-decoder structure combined with the object-focus network on the multi-layer feature maps of the feature pyramid. Our extensive evaluations of the VisDrone, MS-COCO, and UAVid datasets demonstrate that our model outperforms other DETR-based detectors in both detection accuracy and computational speed, highlighting its superior performance. These results indicate a significant advancement in the field of high-resolution object detection utilizing transformer architecture. [ABSTRACT FROM AUTHOR]

Deep learning models have enabled significant performance improvements to remote sensing image processing. Usually, a large number of training samples is required for detection models. In this study, a dynamic simulation training strategy is designed to generate samples in real time during training. The few adversarial examples are not only directly involved in the training but are also used to fit the distribution model of adversarial noise, helping the real-time generated samples to be similar to adversarial examples. The noise of the training samples is randomly generated according to the distribution model, and the random variation of training inputs reduces the overfitting phenomenon. To enhance the detectability of adversarial noise, the input model is no longer a normalized image but a JPEG error image. Experiments show that with the proposed dynamic simulation training strategy, common classification models such as ResNet and DenseNet can effectively detect adversarial examples.

The problem with the acquisition of satellite imagery in the tropics, especially in the area around the equator is that it is almost covered by clouds throughout the year. Users need cloud cover information and the possibility of obtaining cloudless satellite images before they get the data. An overview of the availability of cloud coverage distribution, especially those presented in a spatial format, was very beneficial and increased efficiency for users to select image data in the area of interest (AoI). This study aimed to develop a cloud removal, so-called blended cloud removal (BCR) model, which was applied in a part of West Java Province. The data used for this study were 33 images of quick looks at high-resolution satellite images of the 2013-2015 period that could be obtained free of charge on the website. The results showed that the distribution of efficiency was that AoI-1 obtained 99.67% efficiency of cloud removal image, AoI-2 was 76.51%, and AoI-3 obtained 98.34%. These three AoI locations have an average efficiency of 91.50%. As a result, there was substantial evidence that fewer than 10% of cloud cover remains after cloud removal. This suggests that by using the BCR model, a considerable change in cloud cover for the AoI location might be obtained, meeting the Geospatial Information Agency's standards. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Remote Sensing is the property of Editorial Office of Journal of Remote Sensing & Science Publishing Co. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Abstract Restoring forest ecosystems has become an increasingly high priority for land managers across the American West. Millions of hectares of forest are in need of drastic yet strategic reductions in density (e.g., basal area). Meeting the restoration and management goals requires quantifying metrics of vertical and horizontal forest structure, which has relied upon field‐based measurements, manned airborne or satellite remote sensing datasets. We used unmanned aerial vehicle (UAV) image‐derived Structure‐from‐Motion (SfM) models and high‐resolution multispectral orthoimagery in this study to quantify vertical and horizontal forest structure at both the fine‐ (

Images of sediment cores are often acquired to preserve primary color information, before such profiles are altered by subsequent sampling and destructive analyses. In many cases, however, no post-processing of these images is undertaken to extract information, despite the fact that image processing can be used to describe and measure structures within the sample. Improvements of RGB (Red/Green/Blue) cameras and image processing algorithms now enable acquisition of high-resolution, metrically calibrated pictures called ortho-images, which have great potential. The way to obtain such ortho-images is by processing several raw images. We propose a semi-automated method that uses metrically calibrated targets to create the ortho-image, using Agisoft Photoscan software and a Python script. The method was tested on sediment cores up to 1.5 m long. It was compared to an approach without markers, one that uses only image matching. The proposed method showed better resolution and less distortion (GSD: 59 µm, RMSE: 7–18 µm). Images acquired without calibrated targets can still be used, by manually positioning points that can then be metrically calibrated. This approach is very useful for smartphone images taken in the field. There are many potential applications for such images of sediment cores, for instance as metric stratigraphic logs to facilitate description of the profile by unit, to study and measure structures (e.g. laminae, instantaneous deposits), or use of image registration or data fusion to create spatial landmarks for non-destructive sensors or destructive laboratory analyses. High-resolution metrically calibrated RGB images of sediment cores are simple to acquire and can play an important role in paleoclimate and paleoenvironmental studies. [ABSTRACT FROM AUTHOR]

Inspired by recent promising developments by deep learning networks, this letter presents a novel approach based on a multifeature long short-term memory (LSTM) network with an optimal unit number for extracting buildings from multitemporal high-resolution optical satellite imagery. The algorithm first extracts specified multifeature data of buildings, including spectral, shape, texture, and indices from multitemporal high-resolution imagery. Next, a designed LSTM network architecture with an optimal unit number is trained using the multifeature data to extract buildings at the pixel level. The proposed network is compared with popular deep learning-based networks, including VGG, U-Net, and ResNet. Finally, postprocessing (e.g., morphological algorithm) is employed to optimize the classification results produced by deep learning. The approach is validated on a newly created data set using Gaofen-2 (GF-2) imagery that contains various buildings on a campus with different materials, shapes, sizes, heights, and directions. Experimental results indicate that the proposed approach outperforms current deep learning-based methods for building extraction and has a potential for practical applications. [ABSTRACT FROM AUTHOR]

In desert area, the precipitation is less and the evaporation is fast. Groundwater is the main source of oasis development. The balance of oasis environment depends on the rational use of groundwater. Sustainable utilization of water resources and effective irrigation to prevent soil salinization, desertification, and vegetation degradation caused by improper irrigation are directly related to the sustainable development of oasis. In this paper, we take the largest desert oasis as an example. Based on a large number of historical data and high-resolution images, field investigation, sampling, and laboratory analysis were carried out to study the salinization characteristics of oasis soil components. We will analyze the changes of soil salinization and irrigation water quality in the past 50 years of oasis development and provide information about oasis water resources, which will provide scientific basis for management and policy-making and for similar research in other oases. This paper also studies the Wetland Park, which is a combination of environmental benefits, aesthetic value, and social value of the theme park, is an important place for tourism, leisure, entertainment, and popular science education. In the wetland, the beautification of plant landscape directly affects the overall landscape quality of the park. At present, the research on urban wetland plant landscape construction mainly focuses on qualitative suggestions based on subjective evaluation. This study puts forward quantitative suggestions based on the subjective evaluation of establishing relevant system with objective factors as indicators. The research results can provide a theoretical basis for the management of urban wetland park before and after maintenance. In this paper, through the study of high-resolution image, it is applied to oasis soil composition research and park plant landscape design and then to promoting the sustainable development of oasis and plant landscape. [ABSTRACT FROM AUTHOR]

To solve the problem of low accuracy and slow speed of drone detection in high-resolution images with fixed cameras, we propose a detection method combining background difference and lightweight network SAG-YOLOv5s. First, background difference is used to extract potential drone targets in high-resolution images, eliminating most of the background to reduce computational overhead. Secondly, the Ghost module and SimAM attention mechanism are introduced on the basis of YOLOv5s to reduce the total number of model parameters and improve feature extraction, and α-DIoU loss is used to replace the original DIoU loss to improve the accuracy of bounding box regression. Finally, to verify the effectiveness of our method, a high-resolution drone dataset is made based on the public data set. Experimental results show that the detection accuracy of the proposed method reaches 97.6%, 24.3 percentage points higher than that of YOLOv5s, and the detection speed in 4K video reaches 13.2 FPS, which meets the actual demand and is significantly better than similar algorithms. It achieves a good balance between detection accuracy and detection speed and provides a method benchmark for high-resolution drone detection under a fixed camera.

This article presents a dataset of raw microscopic images of the prefrontal cortex from wistar rat tissues, after an induced stroke, stained with NeuN antibody. The raw images were captured using a microscope equipped with a digital camera. The dataset is useful for testing techniques for the improvement, registration, and stitching to generate a high-resolution image with a full reconstruction of tissues. Besides, this dataset can be used to assess the neuronal brain after an ischemic event. The dataset contains 1370 microscope images with 20x magnification and 36 (Hierarchical Data Format version 5) hdf5 files with homography matrices between every pair of sequential images per tissue rows.

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Support vector machine (SVM) is a powerful cognitive and learning algorithm in the domain of pattern recognition and image classification. However, the generalization ability of SVM is limited when processing classification of high-resolution remote sensing images. One chief reason for this is that the Euclidean distance-based distance matrix in traditional SVM treats different samples equally and overlooks the global distribution of samples. To construct a more effective SVM-based classification method, this paper proposes a multi-scale Mahalanobis kernel-based SVM classifier. In this new method, we first introduce a Mahalanobis distance kernel to improve the global cognitive learning ability of SVM. Then, the Mahalanobis distance kernel is embedded to the multi-scale kernel learning (MSKL) to construct a novel multi-scale Mahalanobis kernel, in which the parameters are optimized by a bio-inspired algorithm, named differential evolution. Finally, the new method is extended to the classification of high-resolution remote sensing images based on the spatial-spectral features. The comparison experiments of five public UCI datasets and two high-resolution remote sensing images verify that the Mahalanobis distance-based method can obtain more accurate classification results than that of the Euclidean distance-based method. In addition, the proposed method produced the best classification results in all the experiments. The global cognitive learning ability of Mahalanobis distance-based method is stronger than that of the Euclidean distance-based method. In addition, this study indicates that the optimized MSKL are potential for the interpretation and understanding of complicated high-resolution remote sensing scene. [ABSTRACT FROM AUTHOR]

Road network extraction plays a critical role in data updating, urban development, and decision support. To improve the efficiency of labeling road datasets and addressing the problems of traditional methods of manually extracting road networks from high-resolution images, such as their slow speed and heavy workload, this paper proposes a semi-automatic method of road network extraction from high-resolution remote-sensing images. The proposed method needs only a few points to extract a single road in the image. After the roads are extracted one by one, the road network is generated according to the width of each road and the spatial relationships among the roads. For this purpose, we use regional growth, morphology, vector tracking, vector simplification, endpoint modification, road connections, and intersection connections to generate road networks. Experiments on four images with different terrains and different resolutions show that this method has high extraction accuracy under different image conditions. The comparisons with the semi-automatic GVF-snake method based on regional growth also showed its advantages and potentiality. The proposed method is a novel form of semi-automatic road network extraction, and it significantly increases the efficiency of road network extraction.