Change Detection of Multi-temporal Remote Sensing Data Using Wavelet-Based Fusion and K-Means Clustering

In this paper, the problem of change detection from Multi-temporal remote sensing images is addressed. To that end, we present a measure of the observed change by combining the two related tasks. In their preliminary design stage, a method constructs difference image by wavelet-based fusing the results of differencing operation and log-ratio operation. Then, by taking into account the spatial-Neighborhood information, a change detection algorithm based k-means clustering is developed to obtain quantitative detection results. The experiments bring out the efficiency of the proposed technique to interpret each change.