Multiscale change detection in multitemporal satellite images

Abstract

In this letter, we propose a novel technique for unsupervised change detection in multitemporal satellite images. The difference image which is computed from multitemporal images acquired on the same geographical area at two different time instances is decomposed using S-levels undecimated discrete wavelet transform (UDWT). For each pixel in the difference image, a multiscale feature vector is extracted using the subbands of the UDWT decomposition and the difference image itself. The final change detection map is achieved by clustering the multiscale feature vectors using κ-means algorithm into two disjoint classes: changed and unchanged. Experimental results confirm the efficacy of the proposed approach on both optical and synthetic aperture radar images. © 2009 IEEE.