Query：
学者姓名：李蒙蒙
Refining：
Year
Type
Indexed by
Source
Complex
Former Name
Co-
Language
Clean All
Abstract ：
Building type information indicates the functional properties of buildings and plays a crucial role in smart city development and urban socioeconomic activities. Existing methods for classifying building types often face challenges in accurately distinguishing buildings between types while maintaining well-delineated boundaries, especially in complex urban environments. This study introduces a novel framework, i.e., CNN-Transformer cross-attention feature fusion network (CTCFNet), for building type classification from very high resolution remote sensing images. CTCFNet integrates convolutional neural networks (CNNs) and Transformers using an interactive cross-encoder fusion module that enhances semantic feature learning and improves classification accuracy in complex scenarios. We develop an adaptive collaboration optimization module that applies human visual attention mechanisms to enhance the feature representation of building types and boundaries simultaneously. To address the scarcity of datasets in building type classification, we create two new datasets, i.e., the urban building type (UBT) dataset and the town building type (TBT) dataset, for model evaluation. Extensive experiments on these datasets demonstrate that CTCFNet outperforms popular CNNs, Transformers, and dual-encoder methods in identifying building types across various regions, achieving the highest mean intersection over union of 78.20% and 77.11%, F1 scores of 86.83% and 88.22%, and overall accuracy of 95.07% and 95.73% on the UBT and TBT datasets, respectively. We conclude that CTCFNet effectively addresses the challenges of high interclass similarity and intraclass inconsistency in complex scenes, yielding results with well-delineated building boundaries and accurate building types.
Keyword ：
Accuracy Accuracy Architecture Architecture Buildings Buildings Building type classification Building type classification CNN-transformer networks CNN-transformer networks cross-encoder cross-encoder Earth Earth Feature extraction Feature extraction feature interaction feature interaction Optimization Optimization Remote sensing Remote sensing Semantics Semantics Transformers Transformers very high resolution remote sensing very high resolution remote sensing Visualization Visualization
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Zhang, Shaofeng , Li, Mengmeng , Zhao, Wufan et al. Building Type Classification Using CNN-Transformer Cross-Encoder Adaptive Learning From Very High Resolution Satellite Images [J]. | IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING , 2025 , 18 : 976-994 . |
| MLA | Zhang, Shaofeng et al. "Building Type Classification Using CNN-Transformer Cross-Encoder Adaptive Learning From Very High Resolution Satellite Images" . | IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING 18 (2025) : 976-994 . |
| APA | Zhang, Shaofeng , Li, Mengmeng , Zhao, Wufan , Wang, Xiaoqin , Wu, Qunyong . Building Type Classification Using CNN-Transformer Cross-Encoder Adaptive Learning From Very High Resolution Satellite Images . | IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING , 2025 , 18 , 976-994 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Change detection is a fundamental yet challenging task in remote sensing, crucial for monitoring urban expansion, land use changes, and environmental dynamics. However, compared with common color images, objects in remote sensing images exhibit minimal interclass variation and significant intraclass variation in the spectral dimension, with obvious scale inconsistency in the spatial dimension. Change detection complexity presents significant challenges, including differentiating similar objects, accounting for scale variations, and identifying pseudo changes. This research introduces a dual fine-grained network with a frequency Transformer (named as FTransDF-Net) to address the above issues. Specifically, for small-scale and approximate spectral ground objects, the network employs an encoder-decoder architecture consisting of dual fine-grained gated (DFG) modules. This enables the extraction and fusion of fine-grained level information in dual dimensions of features, facilitating a comprehensive analysis of their differences and correlations. As a result, a dynamic fusion representation of salient information is achieved. Additionally, we develop a lightweight frequency transformer (LFT) with minimal parameters for detecting large-scale ground objects that undergo significant changes over time. This is achieved by incorporating a frequency attention (FA) module, which utilizes Fourier transform to model long-range dependencies and combines global adaptive attentive features with multi-level fine-grained features. Our comparative experiments across four publicly available datasets demonstrate that FTransDF-Net reaches advanced results. Importantly, it outperforms the leading comparison method by 1.23% and 2.46% regarding IoU metrics concerning CDD and DSIFN, respectively. Furthermore, efficacy for each module is substantiated through ablation experiments. The code is accessible on https://github.com/LeeThrzz/FTrans-DF-Net.
Keyword ：
Change detection Change detection Dual fine-grained Dual fine-grained Frequency transformer Frequency transformer Remote sensing Remote sensing
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Li, Zhen , Zhang, Zhenxin , Li, Mengmeng et al. Dual Fine-Grained network with frequency Transformer for change detection on remote sensing images [J]. | INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION , 2025 , 136 . |
| MLA | Li, Zhen et al. "Dual Fine-Grained network with frequency Transformer for change detection on remote sensing images" . | INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION 136 (2025) . |
| APA | Li, Zhen , Zhang, Zhenxin , Li, Mengmeng , Zhang, Liqiang , Peng, Xueli , He, Rixing et al. Dual Fine-Grained network with frequency Transformer for change detection on remote sensing images . | INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION , 2025 , 136 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Heterogeneous change detection is a task of considerable practical importance and significant challenge in remote sensing. Heterogeneous change detection involves identifying change areas using remote sensing images obtained from different sensors or imaging conditions. Recently, research has focused on feature space translation methods based on deep learning technology for heterogeneous images. However, these types of methods often lead to the loss of original image information, and the translated features cannot be efficiently compared, further limiting the accuracy of change detection. For these issues, we propose a cross-modal feature interaction network (CMFINet). Specifically, CMFINet introduces a cross-modal interaction module (CMIM), which facilitates the interaction between heterogeneous features through attention exchange. This approach promotes consistent representation of heterogeneous features while preserving image characteristics. Additionally, we design a differential feature extraction module (DFEM) to enhance the extraction of true change features from spatial and channel dimensions, facilitating efficient comparison after feature interaction. Extensive experiments conducted on the California, Toulouse, and Wuhan datasets demonstrate that CMFINet outperforms eight existing methods in identifying change areas in different scenes from multimodal images. Compared to the existing methods applied to the three datasets, CMFINet achieved the highest F1 scores of 83.93%, 75.65%, and 95.42%, and the highest mIoU values of 85.38%, 78.34%, and 94.87%, respectively. The results demonstrate the effectiveness and applicability of CMFINet in heterogeneous change detection.
Keyword ：
attention mechanisms attention mechanisms Change detection Change detection CNN CNN feature interaction feature interaction heterogeneous remote sensing images heterogeneous remote sensing images
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Yang, Zhiwei , Wang, Xiaoqin , Lin, Haihan et al. Cross-modal feature interaction network for heterogeneous change detection [J]. | GEO-SPATIAL INFORMATION SCIENCE , 2025 . |
| MLA | Yang, Zhiwei et al. "Cross-modal feature interaction network for heterogeneous change detection" . | GEO-SPATIAL INFORMATION SCIENCE (2025) . |
| APA | Yang, Zhiwei , Wang, Xiaoqin , Lin, Haihan , Li, Mengmeng , Lin, Mengjing . Cross-modal feature interaction network for heterogeneous change detection . | GEO-SPATIAL INFORMATION SCIENCE , 2025 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Building extraction from very high-resolution remote-sensing images still faces two main issues: (1) small buildings are severely omitted and the extracted building shapes have a low consistency with ground truths. (2) supervised deep-learning methods have poor performance in few-shot scenarios, limiting the practical application of these methods. To address the first issue, we propose an asymmetric Siamese multitask network integrating adversarial edge learning called ASMBR-Net for building extraction. It contains an efficient asymmetric Siamese feature extractor comprising pre-trained backbones of convolutional neural networks and Transformers under pre-training and fine-tuning paradigms. This extractor balances the local and global feature representation and reduces training costs. Adversarial edge-learning technology automatically integrates edge constraints and strengthens the modeling ability of small and complex building-shaped patterns. Aiming to overcome the second issue, we introduce a self-training framework and design an instance transfer strategy to generate reliable pseudo-samples. We examined the proposed method on the WHU and Massachusetts (MA) datasets and a self-constructed Dongying (DY) dataset, comparing it with state-of-the-art methods. The experimental results show that our method achieves the highest F1-score of 96.06%, 86.90%, and 84.98% on the WHU, MA, and DY datasets, respectively. Ablation experiments further verify the effectiveness of the proposed method. The code is available at: https://github.com/liuxuanguang/ASMBR-Net
Keyword ：
Adversarial learning Adversarial learning Building extraction Building extraction Multitask learning Multitask learning Self-learning Self-learning VHR remote-sensing image VHR remote-sensing image
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Liu, Xuanguang , Li, Yujie , Dai, Chenguang et al. Extraction buildings from very high-resolution images with asymmetric siamese multitask networks and adversarial edge learning [J]. | INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION , 2025 , 136 . |
| MLA | Liu, Xuanguang et al. "Extraction buildings from very high-resolution images with asymmetric siamese multitask networks and adversarial edge learning" . | INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION 136 (2025) . |
| APA | Liu, Xuanguang , Li, Yujie , Dai, Chenguang , Zhang, Zhenchao , Ding, Lei , Li, Mengmeng et al. Extraction buildings from very high-resolution images with asymmetric siamese multitask networks and adversarial edge learning . | INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION , 2025 , 136 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Precise information on agricultural parcels is crucial for effective farm management, crop mapping, and monitoring. Current techniques often encounter difficulties in automatically delineating vectorized parcels from remote sensing images, especially in irregular-shaped areas, making it challenging to derive closed and vectorized boundaries. To address this, we treat parcel delineation as identifying valid parcel vertices from remote sensing images to generate parcel polygons. We introduce a Point-Line-Region interactive multitask network (PLR-Net) that jointly learns semantic features of parcel vertices, boundaries, and regions through point-, line-, and region-related subtasks within a multitask learning framework. We derived an attraction field map (AFM) to enhance the feature representation of parcel boundaries and improve the detection of parcel regions while maintaining high geometric accuracy. The point-related subtask focuses on learning features of parcel vertices to obtain preliminary vertices, which are then refined based on detected boundary pixels to derive valid parcel vertices for polygon generation. We designed a spatial and channel excitation module for feature interaction to enhance interactions between points, lines, and regions. Finally, the generated parcel polygons are refined using the Douglas-Peucker algorithm to regularize polygon shapes. We evaluated PLR-Net using high-resolution GF-2 satellite images from the Shandong, Xinjiang, and Sichuan provinces of China and medium-resolution Sentinel-2 images from The Netherlands. Results showed that our method outperformed existing state-of-the-art techniques (e.g., BsiNet, SEANet, and Hisup) in pixel- and object-based geometric accuracy across all datasets, achieving the highest IoU and polygonal average precision on GF2 datasets (e.g., 90.84% and 82.00% in Xinjiang) and on the Sentinel-2 dataset (75.86% and 47.1%). Moreover, when trained on the Xinjiang dataset, the model successfully transferred to the Shandong dataset, achieving an IoU score of 83.98%. These results demonstrate that PLR-Net is an accurate, robust, and transferable method suitable for extracting vectorized parcels from diverse regions and types of remote sensing images. The source codes of our model are available at https://github.com/mengmengli01/PLR-Net-demo/tree/main.
Keyword ：
Agricultural parcel delineation Agricultural parcel delineation Multitask neural networks Multitask neural networks PLR-Net PLR-Net Point-line-region interactive Point-line-region interactive Vectorized parcels Vectorized parcels
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Li, Mengmeng , Lu, Chengwen , Lin, Mengjing et al. Extracting vectorized agricultural parcels from high-resolution satellite images using a Point-Line-Region interactive multitask model [J]. | COMPUTERS AND ELECTRONICS IN AGRICULTURE , 2025 , 231 . |
| MLA | Li, Mengmeng et al. "Extracting vectorized agricultural parcels from high-resolution satellite images using a Point-Line-Region interactive multitask model" . | COMPUTERS AND ELECTRONICS IN AGRICULTURE 231 (2025) . |
| APA | Li, Mengmeng , Lu, Chengwen , Lin, Mengjing , Xiu, Xiaolong , Long, Jiang , Wang, Xiaoqin . Extracting vectorized agricultural parcels from high-resolution satellite images using a Point-Line-Region interactive multitask model . | COMPUTERS AND ELECTRONICS IN AGRICULTURE , 2025 , 231 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Tea trees (Camellia sinensis), a quintessential homestead agroforestry crop cultivated in over 60 countries, hold significant economic and social importance as a vital specialty cash crop. Accurate nationwide crop data is imperative for effective agricultural management and resource regulation. However, many regions grapple with a lack of agroforestry cash crop data, impeding sustainable development and poverty eradication, especially in economically underdeveloped countries. The large-scale mapping of tea plantations faces substantial limitations and challenges due to their sparse distribution compared to field crops, unfamiliar characteristics, and spectral confusion among various land cover types (e.g., forests, orchards, and farmlands). To address these challenges, we developed the Manual management And Phenolics substance-based Tea mapping (MAP-Tea) framework by harnessing Sentinel-1/2 time series images for automated tea plantation mapping. Tea trees, exhibiting higher phenolic content, evergreen characteristics, and multiple shoot sprouting, result in extensive canopy coverage, stable soil exposure, and radar backscatter signal interference from frequent picking activities. We developed three phenology-based indicators focusing on phenolic content, vegetation coverage, and canopy texture leveraging the temporal features of vegetation, pigments, soil, and radar backscattering. Characteristics of biochemical substance content and manual management measures were applied to tea mapping for the first time. The MAP-Tea framework successfully generated China's first updated 10 m resolution tea plantation map in 2022. It achieved an overall accuracy of 94.87% based on 16,712 reference samples, with a kappa coefficient of 0.83 and an F1 score of 85.63%. The tea trees are typically cultivated in mountainous and hilly areas with a relatively low planting density (averaging about 10%). Alpine tea trees exhibited a notably dense concentration and dominance, mainly found in regions with elevations ranging from 700 m to 2000 m and slopes between 2 degrees to 18 degrees. The areas with low altitudes and slopes hold the largest tea plantation area and output. As the slope increased, there was a gradual decline in the dominance of tea areas. The results suggest a good potential for the knowledge-based approaches, combining biochemical substance content and human activities, for national-scale tea plantation mapping in complex environment conditions and challenging landscapes, providing important reference significance for mapping other agroforestry crops. This study contributes significantly to advancing the achievement of the Sustainable Development Goals (SDGs) considering the crucial role that agroforestry crops play in fostering economic growth and alleviating poverty. The first 10m national Tea tree data products in China with good accuracy (ChinaTea10m) are publicly accessed at https://doi.org/10.6084/m9.figshare .25047308.
Keyword ：
Agroforestry crop mapping Agroforestry crop mapping Phenology-based algorithm Phenology-based algorithm Sentinel-1/2 Sentinel-1/2 Special cash crop Special cash crop Tea plantation Tea plantation
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Peng, Yufeng , Qiu, Bingwen , Tang, Zhenghong et al. Where is tea grown in the world: A robust mapping framework for agroforestry crop with knowledge graph and sentinels images [J]. | REMOTE SENSING OF ENVIRONMENT , 2024 , 303 . |
| MLA | Peng, Yufeng et al. "Where is tea grown in the world: A robust mapping framework for agroforestry crop with knowledge graph and sentinels images" . | REMOTE SENSING OF ENVIRONMENT 303 (2024) . |
| APA | Peng, Yufeng , Qiu, Bingwen , Tang, Zhenghong , Xu, Weiming , Yang, Peng , Wu, Wenbin et al. Where is tea grown in the world: A robust mapping framework for agroforestry crop with knowledge graph and sentinels images . | REMOTE SENSING OF ENVIRONMENT , 2024 , 303 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Semantic change detection (SCD) from very high-resolution (VHR) images involves two key challenges: 1) the global features of bitemporal images tend to be extracted insufficiently, leading to imprecise land cover semantic classification results; and 2) the detected changed objects exhibit ambiguous boundaries, resulting in low geometric accuracy. To address these two issues, we propose an SCD method called TBSCD-Net based on a multitask learning framework to simultaneously identify different types of semantic changes and regularize change boundaries. First, we construct a hybrid encoder combining transformer and convolutional neural network (CNN) (TCEncoder) to enhance the extraction of global context information. A bitemporal semantic linkage module (Bi-SLM) is embedded into the TCEncoder to enhance the semantic correlations between bitemporal images. Second, we introduce a boundary-region joint extractor based on Laplacian operators (LOBRE) to regularize the changed objects. We evaluated the effectiveness of the proposed method using the SECOND dataset and a Fuzhou GF-2 SCD dataset (FZ-SCD) and compared it with seven existing methods. The proposed method performed better than the other evaluated methods as it achieved 24.42% separation kappa (Sek) and 20.18% global total classification error (GTC) on the SECOND dataset and 23.10% Sek and 23.15% GTC on the FZ-SCD dataset. The results of ablation studies on the FZ-SCD dataset also verified the effectiveness of the developed modules for SCD.
Keyword ：
Boundary regularization Boundary regularization Decoding Decoding Feature extraction Feature extraction Land surface Land surface Laplace equations Laplace equations multitask learning multitask learning semantic change detection (SCD) semantic change detection (SCD) Semantics Semantics Siamese neural network Siamese neural network Task analysis Task analysis Transformers Transformers very high-resolution (VHR) satellite images very high-resolution (VHR) satellite images
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Liu, Xuanguang , Dai, Chenguang , Zhang, Zhenchao et al. TBSCD-Net: A Siamese Multitask Network Integrating Transformers and Boundary Regularization for Semantic Change Detection From VHR Satellite Images [J]. | IEEE GEOSCIENCE AND REMOTE SENSING LETTERS , 2024 , 21 . |
| MLA | Liu, Xuanguang et al. "TBSCD-Net: A Siamese Multitask Network Integrating Transformers and Boundary Regularization for Semantic Change Detection From VHR Satellite Images" . | IEEE GEOSCIENCE AND REMOTE SENSING LETTERS 21 (2024) . |
| APA | Liu, Xuanguang , Dai, Chenguang , Zhang, Zhenchao , Li, Mengmeng , Wang, Hanyun , Ji, Hongliang et al. TBSCD-Net: A Siamese Multitask Network Integrating Transformers and Boundary Regularization for Semantic Change Detection From VHR Satellite Images . | IEEE GEOSCIENCE AND REMOTE SENSING LETTERS , 2024 , 21 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Current land use classification models based on very high-resolution (VHR) remote sensing images often suffer from high sample dependence and poor transferability. To address these challenges, we propose an unsupervised multisource domain adaptation framework for cross-domain land use classification that eliminates the need for repeatedly using source domain data. Our method uses the Swin Transformer as the backbone of the source domain model to extract features from multiple source domain samples. The model is trained on source domain samples, and unlabeled target domain samples are then used for target domain model training. To minimize the feature discrepancies between the source and target domains, we use a weighted information maximization loss and self-supervised pseudolabels to alleviate cross-domain classification noise. We conducted experiments on four public scene datasets and four new land use scene datasets created from different VHR images in four Chinese cities. Results show that our method outperformed three existing single-source cross-domain methods (i.e., DANN, DeepCORAL, and DSAN) and four multisource cross-domain methods (i.e., M3SDA, PTMDA, MFSAN, and SHOT), achieving the highest average classification accuracy and strong stability. We conclude that our method has high potential for practical applications in cross-domain land use classification using VHR images.
Keyword ：
Adaptation models Adaptation models Computational modeling Computational modeling Cross-domain classification Cross-domain classification Data models Data models Feature extraction Feature extraction land use classification land use classification multisource domain adaptation multisource domain adaptation Remote sensing Remote sensing Swin transformer Swin transformer Transformers Transformers Urban areas Urban areas very high resolution (VHR) remote sensing images very high resolution (VHR) remote sensing images
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Li, Mengmeng , Zhang, Congcong , Zhao, Wufan et al. Cross-Domain Urban Land Use Classification via Scenewise Unsupervised Multisource Domain Adaptation With Transformer [J]. | IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING , 2024 , 17 : 10051-10066 . |
| MLA | Li, Mengmeng et al. "Cross-Domain Urban Land Use Classification via Scenewise Unsupervised Multisource Domain Adaptation With Transformer" . | IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING 17 (2024) : 10051-10066 . |
| APA | Li, Mengmeng , Zhang, Congcong , Zhao, Wufan , Zhou, Wen . Cross-Domain Urban Land Use Classification via Scenewise Unsupervised Multisource Domain Adaptation With Transformer . | IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING , 2024 , 17 , 10051-10066 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Extraction of land use information from very high resolution (VHR) images plays a crucial role in urban planning and management. The study aims to extract urban land use information using VHR images and open geographic data using graph neural networks. We first obtained land cover objects using a semantic segmentation model. The spatial topological relationships between land cover objects were then modeled using graph theory and represented as graph-structured data, in which the attributes of graph nodes were computed based upon points of interest (POI) data and classified land cover map. Last, we used graph neural network to learn high-level structural features for urban land use classification. The proposed method was applied to the core urban area of Fuzhou city, China. Results showed that graph neural networks are effective for urban land use classification from VHR images, and integrating open geographic data further improves the accuracy of urban land use classification to 87% compared to the 84%accuracy obtained by using only VHR images. Our method exhibits high potential for extracting fine-grained urban land use in various urban areas. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.
Keyword ：
Classification (of information) Classification (of information) Data integration Data integration Graph neural networks Graph neural networks Graph theory Graph theory Image classification Image classification Image enhancement Image enhancement Information use Information use Land use Land use Remote sensing Remote sensing Semantics Semantics Semantic Web Semantic Web Urban planning Urban planning
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Gai, Xinyi , Li, Mengmeng , Chu, Guozhong et al. Integrating open geographic data for urban land use classification using graph neural networks from high-resolution remote sensing imagery [C] . 2024 . |
| MLA | Gai, Xinyi et al. "Integrating open geographic data for urban land use classification using graph neural networks from high-resolution remote sensing imagery" . (2024) . |
| APA | Gai, Xinyi , Li, Mengmeng , Chu, Guozhong , Lou, Kangkai . Integrating open geographic data for urban land use classification using graph neural networks from high-resolution remote sensing imagery . (2024) . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Building change detection is essential to many applications, such as monitoring of urban areas, land use management, and illegal building detection. It has been seen as an effective means to detect building changes from remote-sensing images. This paper proposes an object-based Siamese neural network, labeled as Obj-SiamNet, to detect building changes from high-resolution remote-sensing images. We combine the advantages of object-based image analysis methods and Siamese neural networks to improve the geometric accuracies of detected boundaries. Moreover, we implement the Obj-SiamNet at multiple segmentation levels and automatically construct a set of fuzzy measures to fuse the obtained results at multi-levels. Furthermore, we use generative adversarial methods to generate target-like training samples from publicly available datasets and construct a relatively sufficient training dataset for the Obj-SiamNet model. Finally, we apply the proposed method into three high-resolution remote-sensing datasets, i.e., a GF-2 image-pair in Fuzhou City, and a GF2 image pair in Pucheng County, and a GF-2—GF-7 image pair in Quanzhou City. We also compare the proposed method with three other existing ones, namely, STANet, ChangeNet, and Siam-NestedUNet. Experimental results show that the proposed method performs better than the other three in terms of detection accuracy. (1) Compared with the detection results from single-scale segmentation, the detection results from multi-scale increases the recall rate by up to 32%, the F1-Score increases by up to 25%, and the Global Total Classification error (GTC) decreases by up to 7%. (2) When the number of available samples is limited, the adopted Generative Adversarial Network (GAN) is able to generate effective target-like samples for diverting samples. Compared with the detection without using GAN-generated samples, the proposed detection increases the recall rate by up to 16%, increases the F1-Score by up to 14%, and decreases GTC by 9%. (3) Compared with other change-detection methods, the proposed method improves the detection accuracies significantly, i.e., the F1-Score increases by up to 23%, and GTC decreases by up to 9%. Moreover, the boundaries of the detected changes by the proposed method have a high consistency with that of ground truth. We conclude that the proposed Obj-SiamNet method has a high potential for building change detection from high-resolution remote-sensing images. © 2024 Science Press. All rights reserved.
Keyword ：
Change detection Change detection Fuzzy sets Fuzzy sets Generative adversarial networks Generative adversarial networks Image enhancement Image enhancement Land use Land use Object detection Object detection Remote sensing Remote sensing
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Liu, Xuanguang , Li, Mengmeng , Wang, Xiaoqin et al. Use of object-based Siamese neural network to build change detection from very high resolution remote-sensing images [J]. | National Remote Sensing Bulletin , 2024 , 28 (2) : 437-454 . |
| MLA | Liu, Xuanguang et al. "Use of object-based Siamese neural network to build change detection from very high resolution remote-sensing images" . | National Remote Sensing Bulletin 28 . 2 (2024) : 437-454 . |
| APA | Liu, Xuanguang , Li, Mengmeng , Wang, Xiaoqin , Zhang, Zhenchao . Use of object-based Siamese neural network to build change detection from very high resolution remote-sensing images . | National Remote Sensing Bulletin , 2024 , 28 (2) , 437-454 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Export
| Results： |
Selected to |
| Format： |
