Loading [MathJax]/extensions/MathMenu.js
A New Building Change Detection Method Based on Cross-Temporal Stereo Matching Using Satellite Stereo Imagery | IEEE Journals & Magazine | IEEE Xplore

A New Building Change Detection Method Based on Cross-Temporal Stereo Matching Using Satellite Stereo Imagery


Abstract:

Change detection between images from different perspectives is a difficult problem in building change detection. At the same time, the quality of digital surface model (D...Show More

Abstract:

Change detection between images from different perspectives is a difficult problem in building change detection. At the same time, the quality of digital surface model (DSM) is very important for building change detection. We propose a new building change detection algorithm based on cross-temporal stereo matching (CTSM) to process images with different viewing angles using two temporal optical spaceborne stereo images. The algorithm matches two images with different temporal points and different viewing angles pixel by pixel to obtain unchanged regions and their elevations. Since the elevation of the unchanged area should be unchanged, we use the elevation obtained by CTSM to correct the existing two temporal DSM. The proposed algorithm can correct the wrong elevation of the unchanged area, recover the matching failure area and fill the occluded area to obtain a more complete and accurate DSM. Applying the refined DSM and unchanged area mask to three existing building change detection algorithms, the performance of all algorithms has been greatly improved. The false alarm rate of the result is greatly reduced while the detection rate of changed buildings keeps almost unchanged.
Published in: IEEE Geoscience and Remote Sensing Letters ( Volume: 20)
Article Sequence Number: 6011305
Date of Publication: 14 September 2023

ISSN Information:

Funding Agency:


References

References is not available for this document.