Image Change Detection Based on the Minimum Mean Square Error

The detection of change is one of the most important tasks in remote sensing analysis. In this paper, a novel unsupervised change detection approach by minimizing the mean square error (MSE) is proposed. The difference image computed by the absolute-valued log ratio of the intensity values of two input images is partitioned into two distinct regions according to the change mask. For each region, the mean square error between its difference image values and the average of its difference image values is calculated. In single-band images, the accurate solution of the change mask with minimum MSE can be obtained in an acceptable time. In multi spectral images, it is considered as a multi-objective optimizations problem. GA is used to obtain the optimal compromised solution. The change detection result of the Florida citrus aerial imagery data is provided. Change error matrix and Kappa coefficient are used to assess the effectiveness of the change detection techniques.