Dengji Zhou
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Dengji Zhou has a total of 32 co-authors. Here are the top co-authors by number of shared publications:
- Guizhou Wang
Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
5 shared publications - Ranyu Yin
Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
5 shared publications - Guojin He
Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
4 shared publications - Tengfei Long
Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
3 shared publications - Bin Luo
School of Computer Science and Technology, Anhui University, Hefei, Anhui, China
2 shared publications Guojin He
Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China
2 shared publications
Latest Publications
- Open Access
- 18 January 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (2) , 324
Wind Turbines Small Object Detection in Remote Sensing Images Based on CGA-YOLO: A Case Study in Shandong Province, China
With the rapid development of high-resolution satellite remote sensing technology, wind turbine detection based on remote sensing imagery has emerged as a crucial research area in renewable energy. However, accurate identification of wind turbines remains challenging due to complex geographical backgrounds and their typical appearance as small objects in images, where limited features and background interference hinder detection performance. To address these issues, this paper proposes CGA-YOLO, a specialized network for detecting small targets in high-resolution remote sensing images, and constructs the SDWT dataset, containing Gaofen-2 imagery covering various terrains in Shandong Province, China. The network incorporates three key enhancements: dynamic convolution improves multi-scale feature representation for precise localization; the Convolutional Block Attention Module (CBAM) enhances feature convergence through channel and spatial attention mechanisms; and GhostBottleneck maintains high-resolution details while strengthening feature channels for small targets. Experimental results demonstrate that CGA-YOLO achieves an F1-score of 0.93 and an mAP50 of 0.938 on the SDWT dataset, and obtains an mAP50 of 0.9033 on both RSOD and VEDAI public datasets. CGA-YOLO establishes its superior accuracy over multiple mainstream detection models under identical experimental conditions, confirming its potential as a reliable technical solution for accurate wind turbine identification in complex environments.
- 18 January 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (2) , 324
Open AccessWind Turbines Small Object Detection in Remote Sensing Images Based on CGA-YOLO: A Case Study in Shandong Province, China
With the rapid development of high-resolution satellite remote sensing technology, wind turbine detection based on remote sensing imagery has emerged as a crucial research area in renewable energy. However, accurate identification of wind turbines remains challenging due to complex geographical backgrounds and their typical appearance as small objects in images, where limited features and background interference hinder detection performance. To address these issues, this paper proposes CGA-YOLO, a specialized network for detecting small targets in high-resolution remote sensing images, and constructs the SDWT dataset, containing Gaofen-2 imagery covering various terrains in Shandong Province, China. The network incorporates three key enhancements: dynamic convolution improves multi-scale feature representation for precise localization; the Convolutional Block Attention Module (CBAM) enhances feature convergence through channel and spatial attention mechanisms; and GhostBottleneck maintains high-resolution details while strengthening feature channels for small targets. Experimental results demonstrate that CGA-YOLO achieves an F1-score of 0.93 and an mAP50 of 0.938 on the SDWT dataset, and obtains an mAP50 of 0.9033 on both RSOD and VEDAI public datasets. CGA-YOLO establishes its superior accuracy over multiple mainstream detection models under identical experimental conditions, confirming its potential as a reliable technical solution for accurate wind turbine identification in complex environments.
- Open Access
- 5 December 2025
- Published by International Society of Global Health in Journal of Global Health
- Vol. 15, 04319
The effects of physical activity on diabetic retinopathy in type 2 diabetes using automated vascular analysis: a cohort study
The American Diabetes Association recommends that adults with diabetes engage in ≥150-minute of moderate-intensity exercise or ≥75-minute of vigorous-intensity exercise weekly, distributed over at least three days [1]. Physical activity (PA) protects patients with diabetes from macrovascular complications and reduces mortality, while also improving glycated haemoglobin, triglycerides, blood pressure, and insulin resistance [2]. Diabetic retinopathy […]
- 5 December 2025
- Published by International Society of Global Health in Journal of Global Health
- Vol. 15, 04319
Open AccessThe effects of physical activity on diabetic retinopathy in type 2 diabetes using automated vascular analysis: a cohort study
The American Diabetes Association recommends that adults with diabetes engage in ≥150-minute of moderate-intensity exercise or ≥75-minute of vigorous-intensity exercise weekly, distributed over at least three days [1]. Physical activity (PA) protects patients with diabetes from macrovascular complications and reduces mortality, while also improving glycated haemoglobin, triglycerides, blood pressure, and insulin resistance [2]. Diabetic retinopathy […]
- Open Access
- 26 August 2022
- Published by MDPI AG in Remote Sensing
- Vol. 14 (17) , 4211
A Hierarchical Information Extraction Method for Large-Scale Centralized Photovoltaic Power Plants Based on Multi-Source Remote Sensing Images
In the context of global sustainable development, solar energy is very widely used. The installed capacity of photovoltaic panels in countries around the world, especially in China, is increasing steadily and rapidly. In order to obtain accurate information about photovoltaic panels and provide data support for the macro-control of the photovoltaic industry, this paper proposed a hierarchical information extraction method, including positioning information and shape information, and carried out photovoltaic panel distribution mapping. This method is suitable for large-scale centralized photovoltaic power plants based on multi-source satellite remote sensing images. This experiment takes the three northwest provinces of China as the research area. First, a deep learning scene classification model, the EfficientNet-B5 model, is used to locate the photovoltaic power plants on 16-m spatial resolution images. This step obtains the area that contains or may contain photovoltaic panels, greatly reducing the study area with an accuracy of 99.97%. Second, a deep learning semantic segmentation model, the U2-Net model, is used to precisely locate photovoltaic panels on 2-m spatial resolution images. This step achieves the exact extraction results of the photovoltaic panels from the area obtained in the previous step, with an accuracy of 97.686%. This paper verifies the superiority of a hierarchical information extraction method in terms of accuracy and efficiency through comparative experiments with DeepLabV3+, U-Net, SegNet, and FCN8s. This meaningful work identified 180 centralized photovoltaic power plants in the study area. Additionally, this method makes full use of the characteristics of different remote sensing data sources. This method can be applied to the rapid extraction of global photovoltaic panels.
- 26 August 2022
- Published by MDPI AG in Remote Sensing
- Vol. 14 (17) , 4211
Open AccessA Hierarchical Information Extraction Method for Large-Scale Centralized Photovoltaic Power Plants Based on Multi-Source Remote Sensing Images
In the context of global sustainable development, solar energy is very widely used. The installed capacity of photovoltaic panels in countries around the world, especially in China, is increasing steadily and rapidly. In order to obtain accurate information about photovoltaic panels and provide data support for the macro-control of the photovoltaic industry, this paper proposed a hierarchical information extraction method, including positioning information and shape information, and carried out photovoltaic panel distribution mapping. This method is suitable for large-scale centralized photovoltaic power plants based on multi-source satellite remote sensing images. This experiment takes the three northwest provinces of China as the research area. First, a deep learning scene classification model, the EfficientNet-B5 model, is used to locate the photovoltaic power plants on 16-m spatial resolution images. This step obtains the area that contains or may contain photovoltaic panels, greatly reducing the study area with an accuracy of 99.97%. Second, a deep learning semantic segmentation model, the U2-Net model, is used to precisely locate photovoltaic panels on 2-m spatial resolution images. This step achieves the exact extraction results of the photovoltaic panels from the area obtained in the previous step, with an accuracy of 97.686%. This paper verifies the superiority of a hierarchical information extraction method in terms of accuracy and efficiency through comparative experiments with DeepLabV3+, U-Net, SegNet, and FCN8s. This meaningful work identified 180 centralized photovoltaic power plants in the study area. Additionally, this method makes full use of the characteristics of different remote sensing data sources. This method can be applied to the rapid extraction of global photovoltaic panels.
- 15 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 19 (1545598X) , 1-5
Automatic Framework of Mapping Impervious Surface Growth With Long-Term Landsat Imagery Based on Temporal Deep Learning Model
The impervious surface (IS) cover and its dynamics are key parameters in research about urban and ecology. This letter proposed an automatic framework to map the IS growth end-to-end based on the temporal deep learning (DL) model and long time-series Landsat imagery. First, the training and validating datasets were auto-generated by a joint strategy. Then, a DL network was designed, and the IS growth was predicted in temporal windows. Finally, the results from multi-temporal windows are combined to generate the IS growth map. The data around the core of Beijing, China, is tested, and the result shows that the proposed method could: 1) efficiently model the IS growth; 2) map IS growth with less salt-and-pepper noise and false alarm compared to existing products; and 3) be extended to future data easily.
- 15 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 19 (1545598X) , 1-5
Automatic Framework of Mapping Impervious Surface Growth With Long-Term Landsat Imagery Based on Temporal Deep Learning Model
The impervious surface (IS) cover and its dynamics are key parameters in research about urban and ecology. This letter proposed an automatic framework to map the IS growth end-to-end based on the temporal deep learning (DL) model and long time-series Landsat imagery. First, the training and validating datasets were auto-generated by a joint strategy. Then, a DL network was designed, and the IS growth was predicted in temporal windows. Finally, the results from multi-temporal windows are combined to generate the IS growth map. The data around the core of Beijing, China, is tested, and the result shows that the proposed method could: 1) efficiently model the IS growth; 2) map IS growth with less salt-and-pepper noise and false alarm compared to existing products; and 3) be extended to future data easily.
- Open Access
- 27 October 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 14 (19391404) , 11530-11545
A Large-Scale Mapping Scheme for Urban Building From Gaofen-2 Images Using Deep Learning and Hierarchical Approach
Urban buildings are essential components of cities and an indispensable source of urban geographic information. While there are many research efforts focused on urban buildings extraction, there are few studies on large-scale urban building mapping based on satellite images. In this research, a large-scale urban building mapping scheme based on Gaofen-2 satellite (GF-2) images is proposed based on a hierarchical approach. In this hierarchical approach, urban buildings are regarded as a mixture of dense low-rise buildings (DLBs) and sparse independent buildings (SIBs) stacked in space, which are extracted by a semantic segmentation model and an instance segmentation model, respectively. In this study, GF-2 images and OpenStreetMap data were used to extract DLB using $U^2$ -Net with focal loss. GF-2 images were used to extract SIB using an improved CenterMask model with a deformable convolution network and a spatial coordinate attention module. The main urban area within the 5th ring road of Beijing was selected as the study area. With the trained model, the GF-2 image tiles of Beijing input into the models to first derive coarse maps of DLB and SIB. Postprocessing optimization was performed after combining the maps. The accuracy assessment shows that the overall accuracy of large-scale urban building mapping using the hierarchical approach proposed in this article reaches 91.5%, which is 4.8% higher than that with a traditional method. Overall, the hierarchical approach proposed in this article is effective in large-scale urban building mapping and provides new application opportunities.
- 27 October 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 14 (19391404) , 11530-11545
Open AccessA Large-Scale Mapping Scheme for Urban Building From Gaofen-2 Images Using Deep Learning and Hierarchical Approach
Urban buildings are essential components of cities and an indispensable source of urban geographic information. While there are many research efforts focused on urban buildings extraction, there are few studies on large-scale urban building mapping based on satellite images. In this research, a large-scale urban building mapping scheme based on Gaofen-2 satellite (GF-2) images is proposed based on a hierarchical approach. In this hierarchical approach, urban buildings are regarded as a mixture of dense low-rise buildings (DLBs) and sparse independent buildings (SIBs) stacked in space, which are extracted by a semantic segmentation model and an instance segmentation model, respectively. In this study, GF-2 images and OpenStreetMap data were used to extract DLB using $U^2$ -Net with focal loss. GF-2 images were used to extract SIB using an improved CenterMask model with a deformable convolution network and a spatial coordinate attention module. The main urban area within the 5th ring road of Beijing was selected as the study area. With the trained model, the GF-2 image tiles of Beijing input into the models to first derive coarse maps of DLB and SIB. Postprocessing optimization was performed after combining the maps. The accuracy assessment shows that the overall accuracy of large-scale urban building mapping using the hierarchical approach proposed in this article reaches 91.5%, which is 4.8% higher than that with a traditional method. Overall, the hierarchical approach proposed in this article is effective in large-scale urban building mapping and provides new application opportunities.
- Open Access
Robust Building Extraction for High Spatial Resolution Remote Sensing Images with Self-Attention Network
Building extraction from high spatial resolution remote sensing images is a hot spot in the field of remote sensing applications and computer vision. This paper presents a semantic segmentation model, which is a supervised method, named Pyramid Self-Attention Network (PISANet). Its structure is simple, because it contains only two parts: one is the backbone of the network, which is used to learn the local features (short distance context information around the pixel) of buildings from the image; the other part is the pyramid self-attention module, which is used to obtain the global features (long distance context information with other pixels in the image) and the comprehensive features (includes color, texture, geometric and high-level semantic feature) of the building. The network is an end-to-end approach. In the training stage, the input is the remote sensing image and corresponding label, and the output is probability map (the probability that each pixel is or is not building). In the prediction stage, the input is the remote sensing image, and the output is the extraction result of the building. The complexity of the network structure was reduced so that it is easy to implement. The proposed PISANet was tested on two datasets. The result shows that the overall accuracy reached 94.50 and 96.15%, the intersection-over-union reached 77.45 and 87.97%, and F1 index reached 87.27 and 93.55%, respectively. In experiments on different datasets, PISANet obtained high overall accuracy, low error rate and improved integrity of individual buildings.
Open AccessRobust Building Extraction for High Spatial Resolution Remote Sensing Images with Self-Attention Network
Building extraction from high spatial resolution remote sensing images is a hot spot in the field of remote sensing applications and computer vision. This paper presents a semantic segmentation model, which is a supervised method, named Pyramid Self-Attention Network (PISANet). Its structure is simple, because it contains only two parts: one is the backbone of the network, which is used to learn the local features (short distance context information around the pixel) of buildings from the image; the other part is the pyramid self-attention module, which is used to obtain the global features (long distance context information with other pixels in the image) and the comprehensive features (includes color, texture, geometric and high-level semantic feature) of the building. The network is an end-to-end approach. In the training stage, the input is the remote sensing image and corresponding label, and the output is probability map (the probability that each pixel is or is not building). In the prediction stage, the input is the remote sensing image, and the output is the extraction result of the building. The complexity of the network structure was reduced so that it is easy to implement. The proposed PISANet was tested on two datasets. The result shows that the overall accuracy reached 94.50 and 96.15%, the intersection-over-union reached 77.45 and 87.97%, and F1 index reached 87.27 and 93.55%, respectively. In experiments on different datasets, PISANet obtained high overall accuracy, low error rate and improved integrity of individual buildings.
- Open Access
- 1 May 2020
- Published by IOP Publishing in IOP Conference Series: Earth and Environmental Science
- Vol. 502 (1)
Intelligently Extracting and Analysing High-Resolution Remote Sensing Images for Demolition and Reclamation Land Assessment in Population Relocation for Poverty Alleviation
In recent years, China has witnessed a rapid development in the field of high-resolution satellite data acquisition. These remote sensing images obtained from satellites have been extensively applied in the social and economic sectors. Population relocation for poverty alleviation is a key task during the implementation of the 13th Five-Year Plan, involving 1,400 counties in 22 provinces in China. This study focuses on the utilisation of high-resolution remote sensing data and data of the national relocation information system to intelligently analyse changes in demolished and reclaimed lands in the selected area. The area is a county in Shandong Province, and according to the data from the national relocation information system, it constitutes 12 centralised resettlement sites hosting 2,510 families comprising 7,708 people, all classified as poverty-stricken households by the state. This method is suitable for the relocation of the entire village areas rather than the scattered relocation of individual households. It also provides an effective support for the local government to save and intensively use land resources, thereby avoiding waste of land resources and construction funds.
- 1 May 2020
- Published by IOP Publishing in IOP Conference Series: Earth and Environmental Science
- Vol. 502 (1)
Open AccessIntelligently Extracting and Analysing High-Resolution Remote Sensing Images for Demolition and Reclamation Land Assessment in Population Relocation for Poverty Alleviation
In recent years, China has witnessed a rapid development in the field of high-resolution satellite data acquisition. These remote sensing images obtained from satellites have been extensively applied in the social and economic sectors. Population relocation for poverty alleviation is a key task during the implementation of the 13th Five-Year Plan, involving 1,400 counties in 22 provinces in China. This study focuses on the utilisation of high-resolution remote sensing data and data of the national relocation information system to intelligently analyse changes in demolished and reclaimed lands in the selected area. The area is a county in Shandong Province, and according to the data from the national relocation information system, it constitutes 12 centralised resettlement sites hosting 2,510 families comprising 7,708 people, all classified as poverty-stricken households by the state. This method is suitable for the relocation of the entire village areas rather than the scattered relocation of individual households. It also provides an effective support for the local government to save and intensively use land resources, thereby avoiding waste of land resources and construction funds.