Multiscale Unsupervised Change Detection on Optical Images Based on Markov Random Fields and Wavelets

In the context of environmental monitoring and natural or industrial disaster management, change-detection methods represent powerful tools for studying and mapping the evolution of the Earth’s surface. In order to optimize the accuracy of the change maps, a multiscale approach can be adopted, that jointly exploits observations at coarser scales (to globally identify changed areas and gain robustness to noise) and finer scales (to improve the detection of details). In this paper, a multiscale contextual unsupervised change-detection method is proposed for optical images. It is based on discrete wavelet transforms and Markov random fields. The imagedifferencing approach to change-detection with optical data is adopted, wavelets are applied to the difference image to extract multiscale features, and Markovian data fusion is used to integrate both these features and the spatial contextual information in the change-detection process. Expectationmaximization and Besag’s algorithms are used to estimate the model parameters. Experiments on real optical images point out the effectiveness of the method, also as compared with state-of-the-art techniques.