Unsupervised change detection for remote sensing images using multiscale decomposition and treelet fusion: A level set approach

With the development of remote sensing technology, using remote sensing images to do change detection is gaining more and more attention. In this paper, a novel approach for unsupervised change detection is proposed. Firstly, SWT is used for multi-scale decomposition; then use treelet to fusion the reconstructed difference images at every level; finally, the change detection map is obtained through a level set segmentation. The effectiveness of our method is validated by experiments on five different data sets from different resources. By comparing with other state-of-art technologies, the results show that the proposed method yields better performance in most cases.