Related references
Note: Only part of the references are listed.
Article
Engineering, Electrical & Electronic
Lei Song et al.
Summary: This work proposes a bibranch fusion network based on axial cross attention (ACABFNet) for change detection research. It effectively combines local and global features extracted by a CNN branch and a transformer branch, respectively, using a bidirectional fusion approach. The experimental results show that ACABFNet outperforms existing change detection algorithms on three datasets.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2023)
Article
Remote Sensing
Zhanming Ma et al.
Summary: Extracting mask information of buildings and water areas from high resolution remote sensing images is important for urban development monitoring and management. Existing semantic segmentation methods lack attention to channel information and suffer from loss of details in complex scenes, leading to misjudgement of buildings and waters. To address these issues, a feature enhancement network (FENet) is proposed to reduce misjudgement and enhance edge contour information by focusing on characteristic information and encoding context information.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2023)
Article
Environmental Sciences
Kai Hu et al.
Summary: A multi-branch convolutional attention network (MCANet) is proposed for accurate segmentation of cloud/snow regions in remote sensing imagery. The network utilizes a double-branch structure to extract spatial and semantic information, improving feature extraction capability. A fusion module is suggested to correctly merge feature information from multiple branches, and a new decoder module is constructed to enhance information recovery and refine segmentation boundaries.
Article
Environmental Sciences
Chao Zhang et al.
Summary: A new cloud detection method is proposed in this paper, which can accurately and efficiently detect smaller clouds and obtain finer edge segmentation. By using ResNet-18 as the backbone, and combining the Multi-scale Global Attention Module and Strip Pyramid Channel Attention Module, the detection accuracy of clouds is improved. The Hierarchical Feature Aggregation Module fuses high-dimensional and low-dimensional features, and the final segmentation effect is obtained by layer-by-layer upsampling. The proposed model achieves excellent results on the Cloud and Cloud Shadow Dataset and the public dataset CSWV.
Article
Environmental Sciences
Dehao Wang et al.
Summary: This article proposes a multi-branch collaborative change-detection network based on Siamese structure (MHCNet). It constructs three branches, the difference branch, global branch, and similar branch, to refine and extract semantic information from remote-sensing images. Four modules are also introduced to assist in better semantic information extraction.
Article
Environmental Sciences
Jing Chen et al.
Summary: The segmentation algorithm for buildings and waters is crucial for efficient land resource planning and utilization. However, current convolutional neural networks (CNNs) in remote sensing images lack sensitivity to spatial position information, resulting in loss of location and edge details and decreased segmentation accuracy. To address these issues, this research proposes a double-branch parallel interactive network that fully utilizes the interactivity of global information in a Swin Transformer network and integrates CNN to capture deeper information. By building a cross-scale multi-level fusion module, the model combines features from CNNs and Swin Transformer to successfully extract semantic information of spatial information and context. An upsampling module for multi-scale fusion is also suggested to recover high-resolution pixel-level features. Experimental results show that the proposed networks maximize the benefits of both models and improve the precision of semantic segmentation of buildings and waters.
Article
Green & Sustainable Science & Technology
Zhanming Ma et al.
Summary: This paper proposes a remote sensing image building and water segmentation method based on a local feature search network (DFSNet). By paying attention to the local feature information between horizontal direction and position, it reduces the misjudgment of buildings and loss of local information of water bodies. The introduced discarding attention module (DAM) and slice pooling module (SPM) address the problems of misjudgment and edge blurring, while the fusion attention up sampling module (FAUM) provides better pixel level attention and more detailed segmentation output. Experimental results show a 2.89% improvement on MIoU indicator, with a final MIoU of 83.73%.
Article
Computer Science, Information Systems
Huiwen Ji et al.
Summary: This paper proposes a multi-supervised feature fusion attention network for cloud and cloud shadow detection in remote sensing imagery. By using multi-scale convolution, channel attention, and spatial attention, the network effectively fuses multi-scale features and improves boundary detection accuracy. The proposed model achieves excellent performance on a self-made dataset, with a mean intersection over union (MIoU) of 94.10%, which is 0.44% higher than other models, and also shows good generalization capability on other datasets.
ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION
(2023)
Article
Automation & Control Systems
Chong Ma et al.
Summary: This paper proposes a dual-branch network for change detection, which mitigates detection errors, omissions, and obtains sharper edges through the use of multi-scale strip convolution module, spatial attention module, and feature fusion network. Experimental results demonstrate that the proposed method outperforms other algorithms.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Environmental Sciences
Nan Jiang et al.
Article
Engineering, Electrical & Electronic
Liguo Weng et al.
Summary: With the improvement of land cover classification in high-definition satellite maps, the traditional methods face difficulties in processing rich details, distorted images, and lack of generalization. To address these issues, a dual-branch parallel network structure called Sgformer is proposed, which enhances the performance of the transformer in the context of high-definition remote sensing maps. Experimental results show that this methodology achieves higher accuracy.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Kai Chen et al.
Summary: Cloud and cloud shadow segmentation is a crucial problem in remote sensing image processing. Traditional deep learning networks typically suffer from the interference of ground objects, noise, and the lack of generalization ability, resulting in inaccurate segmentation and detection of cloud and cloud shadow boundaries. To address these issues, a multiscale attention feature aggregation network is proposed, which leverages residual network to extract semantic information at various levels. This network includes a multiscale strip pooling attention module, deep multihead feedforward transfer attention module, bilateral feature fusion module, and a boundary refinement boosting model. Experimental results demonstrate that the proposed model achieves superior accuracy in complex cloud cover scenes, making it of great significance for cloud detection-related work.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Xuebo Zhang et al.
Summary: Before using imaging algorithms based on monostatic synthetic aperture sonar (SAS), a phase center approximation (PCA) is used to transform multireceiver SAS data to monostatic SAS equivalent signal. However, the space variance of phase error poses a challenge to this conversion. This study presents a generalized PCA model that considers the space variance and compensates for the phase error using sub-block processing.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2023)
Article
Geochemistry & Geophysics
Xin Dai et al.
Summary: Traditional building and water segmentation methods are prone to noise interference and suffer from missed and false detections. Excessive deep learning downsampling results in the loss of feature map information and image location offset, diminishing the overall effectiveness. To address these challenges, a multiscale location attention network (MSLANet) is proposed, which focuses on location-spatial and channel information for accurate edge detail segmentation. The network includes a location channel attention (LCA) unit and a dual-branch multiscale aggregation (DBMSA) unit to improve segmentation accuracy. Experimental results demonstrate that the proposed approach outperforms existing methodologies.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Remote Sensing
Shoukuan Miao et al.
Summary: In the application of remote sensing, cloud blocking poses challenges to the analysis of surface and atmospheric parameters. Existing deep learning methods struggle to accurately segment the edge information of clouds and cloud shadows. To address these issues, a multilevel feature enhanced network is proposed for cloud/shadow segmentation. Experimental results demonstrate the superiority of this method over existing approaches.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2022)
Article
Environmental Sciences
Kai Hu et al.
Summary: This study proposes a multi-scale feature aggregation network for water area segmentation. By designing a deep feature extraction module and a multi-branch aggregation module, it accurately identifies small tributaries in water area images and extracts deep semantic information, achieving improved segmentation accuracy.
Article
Environmental Sciences
Jiahong Gao et al.
Summary: The use of remote sensing images for land cover analysis has great potential. To improve accuracy, a multichannel feature fusion lozenge network is proposed, which utilizes three branches for feature sampling, contextual information, and feature integration, resulting in significantly improved accuracy.
JOURNAL OF APPLIED REMOTE SENSING
(2022)
Article
Computer Science, Artificial Intelligence
Shengguang Chu et al.
Summary: Remote sensing image change detection is a crucial technology for analyzing remote sensing data. Existing mainstream solutions can be categorized into supervised and unsupervised methods. Among supervised methods, deep learning-based remote sensing image change methods are mainly focused on semantic segmentation. However, these methods only utilize deep learning models to extract and process single-level image information without separately training the changed and unchanged areas based on the task characteristics of change detection. Additionally, these models do not integrate deep and shallow-level feature information, resulting in missing local change details. In this paper, a multilateral feature-guided aggregation network is proposed to address these limitations. The network extracts whole information, changed area information, and unchanged area information of remote sensing images through a main network and two auxiliary networks: difference network and assimilation network. The feature aggregation network then aggregates the whole information of the image with the changed and unchanged region information (all subnetworks are composed of convolutional neural networks (CNNs)). The proposed method is end-to-end trainable, and each component in the network does not need to be trained separately.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Mathematics
Kai Hu et al.
Summary: This paper proposes a Federated Learning-Based Graph Convolutional Network (FedGCN) to process irregular data and enhance data privacy protection. With a federated aggregation algorithm based on an online adjustable attention mechanism, the damage caused by inefficient local model parameters to the global model is reduced, improving the fault tolerance and accuracy of the algorithm.
Article
Remote Sensing
Bingyu Chen et al.
Summary: With the improvement of segmentation effect, some studies have been applied to high-resolution remote sensing images. Existing algorithms are limited by short-range context and cannot recover high-resolution details. This paper proposes a multi-level aggregation network (MANet) to optimize the segmentation results by using a global dependency module and a two-path feature refining module.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2022)
Article
Engineering, Electrical & Electronic
Kai Hu et al.
Summary: In this paper, a light-weighted cloud detection network (LCDNet) based on deep learning method is proposed. The network can complete the task of high-precision segmentation with less parameters and computation. The experimental results show that the model shows an excellent performance on cloud and cloud shadow datasets and has excellent generalization performance.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Kecheng Chen et al.
Summary: This article proposes a novel denoising framework for TEM signals using deep convolutional neural networks, which transforms the denoising task into an image denoising task. The framework includes a new signal-to-image transformation method and a deep CNN-based denoiser, achieving better performance.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Computer Science, Interdisciplinary Applications
Yi Qu et al.
Summary: A new method for automatic segmentation of clouds and their shadows in remote sensing images is proposed, using a strip pooling residual network to obtain more accurate local position information and improving segmentation accuracy by combining channel attention and spatial attention.
COMPUTERS & GEOSCIENCES
(2021)
Article
Engineering, Electrical & Electronic
Jie Chen et al.
Summary: Change detection is a fundamental task in remote sensing image processing. While deep learning has provided new tools for this task, current methods lack resistance to pseudochange information. To address this issue, a new method called dual attentive fully convolutional Siamese networks is proposed, achieving significant improvements compared to baseline methods according to experimental results.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2021)
Article
Environmental Sciences
Hao Chen et al.
Article
Environmental Sciences
Junhao Qian et al.
Article
Geochemistry & Geophysics
Shunping Ji et al.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2019)
Article
Geochemistry & Geophysics
Mengya Zhang et al.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2019)
Article
Environmental Sciences
Daifeng Peng et al.
Article
Computer Science, Information Systems
Ling Ke et al.
Review
Environmental Sciences
Zhang Zengxiang et al.
CHINESE GEOGRAPHICAL SCIENCE
(2018)
Article
Remote Sensing
Qing Wang et al.
REMOTE SENSING LETTERS
(2018)
Article
Geochemistry & Geophysics
Yang Zhan et al.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2017)
Article
Chemistry, Analytical
Xin Zhang et al.
Review
Computer Science, Interdisciplinary Applications
Shaohua Zhao et al.
EARTH SCIENCE INFORMATICS
(2017)
Article
Computer Science, Artificial Intelligence
Maoguo Gong et al.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2016)
Article
Geochemistry & Geophysics
Yaoguo Zheng et al.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2014)
Article
Engineering, Electrical & Electronic
Qingxi Tong et al.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2014)
Article
Remote Sensing
Pengfei He et al.
REMOTE SENSING LETTERS
(2014)
Article
Computer Science, Artificial Intelligence
Ludmila I. Kuncheva et al.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2014)
Article
Environmental Sciences
George Xian et al.
REMOTE SENSING OF ENVIRONMENT
(2010)
Article
Remote Sensing
M. A. Torres-Vera et al.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2009)
Article
Remote Sensing
J. S. Deng et al.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2008)
Article
Environmental Sciences
A. Koltunov et al.
REMOTE SENSING OF ENVIRONMENT
(2007)