Implementation of building change detection system for remote sensing images based on deep learning
As to the problems of poor robustness and low efficiency by traditional building change detec-tion methods,a building change detection method for remote sensing images based on deep learning is proposed in this paper.Firstly,DeepLab V3+semantic segmentation is used by the method to extract the building segmentation results from the images in different phases.Secondly,Canny algorithm is used to extract the contours of the binarized segmentation results to construct vector building layers.Lastly,the spatial analysis method is used to extract the difference of the area to label the changed contents.The building change detection system is designed and implemented in this paper to provide an idea for apply-ing the method to practice,and the key technologies concerned are introduced.The feasibility of the DeepLab V3+algorithm to segment buildings and the applicability of the system are proved by a building extraction test.
deep learningsemantic segmentationbuilding extractioncontour extractionconvolution neural network
NETL
NSTL
万方数据
