Acta Geodaetica et Cartographica Sinica ›› 2022, Vol. 51 ›› Issue (12): 2531-2540.doi: 10.11947/j.AGCS.2022.20210177

• Photogrammetry and Remote Sensing • Previous Articles     Next Articles

Correg-YOLOv3: a method for dense buildings detection in high-resolution remote sensing images

CHEN Zhanlong1,2,6, LI Shuangjiang1,7, XU Yongyang1,3,6, XU Daozhu4,5, MA Chao4,5, ZHAO Junli2   

  1. 1. School of Geography and Information Engineering, China University of Geoscience, Wuhan 430078, China;
    2. Key Laboratory of Geological Survey and Evaluation of Ministry of Education, China University of Geosciences, Wuhan 430074, China;
    3. Key Laboratory of Urban Land Resource Monitoring and Simulation, Ministry of Natural Resource, Shenzhen 518034, China;
    4. Xi'an Research Institute of Surveying and Mapping, Xi'an 710054, China;
    5. State Key Laboratory of Geo-Information Engineering, Xi'an 710054, China;
    6. School of Computer Science, China University of Geoscience, Wuhan 430078, China;
    7. Central Southern China Electric Power Design Institute Co., Ltd., Wuhan 430071, China
  • Received:2021-04-07 Revised:2021-10-14 Published:2023-01-12
  • Supported by:
    The National Natural Science Foundation of China(No.41871305);The National Key Research and Development Program of China(No.2017YFC0602204);The Fundamental Research Funds for the Central Universities, China University of Geosciences(Wuhan) (No.CUGQY1945);The Open Fund of Key Laboratory of Geological Survey and Evaluation of Ministry of Education and the Fundamental Research Funds for the Central Universities(No.GLAB2019ZR02);The Open Fund of Laboratory of Urban Land Resources Monitoring and Simulation, Ministry of Natural Resources, China (No.KF-2020-05-068)

Abstract: The exploration of building detection plays an important role in urban planning, smart city and military. Aiming at the problem of high overlapping ratio of detection frames for dense building detection in high-resolution remote sensing images, we present an effective YOLOv3 framework, corner regression-based YOLOv3 (Correg-YOLOv3), to localize dense building accurately. This improved YOLOv3 algorithm establishes a vertex regression mechanism and an additional loss item about building vertex offsets relative to the center point of bounding box. By extending output dimensions, the trained model is able to output the rectangular bounding boxes and the building vertices meanwhile. Finally, we evaluate the performance of the Correg-YOLOv3 on our self-produced data set and provide a comparative analysis qualitatively and quantitatively. The experimental results achieve high performance in precision (96.45%), recall rate (95.75%), F1 score (96.10%) and average precision (98.05%), which were 2.73%, 5.4%, 4.1% and 4.73% higher than that of YOLOv3. Therefore, the proposed algorithm effectively tackles the problem of dense building detection in high-resolution images.

Key words: high resolution remote sensing image, Correg-YOLOv3, corner regression, dense buildings, object detection

CLC Number: