The Method for Change Detection by Remote Sensing Images Based on Comprehensive Weighted Attributes Error Diffusion

摘要

Thanks to the large number of remote sensing images acquired by satellite sensors for the same area over time, it is now feasible to identify potential changes occurred on the ground. In this paper, a comprehensive Weighted Attributes Error Diffusion algorithm (WAED) is proposed to detect changes of remote sensing images. This algorithm dedicated in using multi-attributes to balance effect of each feature that may not be identical in source images from different satellites and in different spacial resolutions. The creativeness of the proposed algorithm lies in two aspects: (1) it combines several attributes to produce general comparing result. (2) The error diffusing process strongly enhances the accuracy of detecting area. Our algorithm contributes to detect change area accurately for remote sensing images from different satellite in different spacial resolution. Experimental results validate the effectiveness of our scheme and demonstrate that it outperforms previous methods.