Change detection based on iterative invariant area histogram matching

Image radiometric distortion is problematic to many remote sensing based image change detection methods. Thus image radiometric normalization is critical to successful change detection and other image analysis. Histogram matching (HM) method has been widely used for radiometric normalization. However, when two corresponding image scenes have partial changes, these changes will introduce additional distortions in the histogram matching process. Obviously, if the change areas in the images can be excluded from histogram matching, the distortions in histogram matching will be avoided. Thus, a novel, iterative change detection based HM normalization algorithm is proposed in this paper. The algorithm of the proposed iterative invariant area histogram matching and an adaptive thresholding schema are first introduced, and the change detection experiment results are then presented. The preliminary results indicate that the proposed method significantly improves the performance of change detection for a given threshold as judged by visual inspection.