Unsupervised image change detection means based on improved 2-D entropy

Change detection in images of a given scene acquired at different times is one of the most interesting topics of remote image processing, which finds important applications within different contexts. Most methods are based on difference image in existing literatures, whose key is how to automatically detect change area of the difference image. In this paper, a novel image change detection means based on 2-D histogram formed by pixel gray levels and the local average gray levels is proposed. First, the 2-D histogram is segmented by a line passing through the diagonal of 2-D histogram in a certain direction. Second, the improved 2-D entropy is defined to acquire the optimal segmentation direction and the optimal threshold by searching 2D maximal entropy. And the 2-D histogram is segmented into unchanged region and change region. Then the change area in different image is detected based on the change region of 2-D histogram. Finally, the change detection means proposed in the paper is compared to traditional means by carrying out an experiment on a synthetic data set artificially generated. Theoretical analysis and experiment result show that this algorithm is more accurate on detection precision and faster on detection speed.