ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume V-3-2020
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., V-3-2020, 45–51, 2020
https://doi.org/10.5194/isprs-annals-V-3-2020-45-2020
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., V-3-2020, 45–51, 2020
https://doi.org/10.5194/isprs-annals-V-3-2020-45-2020

  03 Aug 2020

03 Aug 2020

EARTHQUAKE-DAMAGED REGIONS DETECTION FROM HIGH RESOLUTION IMAGE BASED ON SUPER-PIXEL SEGMENTATION AND DEEP LEARNING

C. Liu1, H. Sui1, Y. Peng1, L. Hua2, and Q. Li3 C. Liu et al.
  • 1State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430079, China
  • 2College of Resources & Environment, Huazhong Agriculture University, Wuhan 430070, China
  • 3Hubei Geomatics Information Center (Hubei Research Institute of Beidou Satellite Navigation Applied Technology), Wuhan 4300779, China

Keywords: Building damage, Super-pixel segmentation, AlexNet, Multi-scale samples, High resolution images

Abstract. Accurate detection and automatic processing of earthquake-damaged regions is essential for effective rescue and post-disaster reconstruction. In this study, we proposed a Combined Super-pixel Segmentation and AlexNet Detection approach (CSSAD) for automatically extracting damaged regions from post-earthquake high-resolution images. Simple Linear Iterative Clustering (SLIC) algorithm was used to segment the high resolution images to obtain more homogeneous geo-objects. Multiscale samples database, which took the different scale effect of damaged regions into account, was constructed based on the geometric centre of each super-pixel. AlexNet, which achieved the automatic extraction of high-level features and accurate identification of target geo-objects, was used to detect the damaged regions. To enhance the localization accuracy, the output of AlexNet was further refined using super-pixel segmentations and masked out of shadow and vegetation. Compared with traditional method, the proposed approach effectively reduces the false and missed detection ratio at least 10 percent.