Polygon Change Detection for SPOT5 Color Image Using Multi-feature-clustering-analysis

This paper applies the multi-feature clustering analysis for SPOT5 color remote sensing images change detection based on image subtraction method and solves the problem of misjudge polygons. Subtraction chart is constructed by computing gray difference between pixels with a quite low threshold; Then, multi-feature-clustering-analysis is utilized to subtraction chart to delete misjudge polygons; Lastly, median filter and morphology algorithm are applied to remove noises caused by image registration and pixel-to-pixel-calculation. Experimental results on SPOT5 images are given to show the validity and advantages of the proposed approach.