Multi-spectral remote sensing images fusion based on k-th central moment

With the remote sense technologies applied into more and more fields, the existing methods of remote sense data fusion, like HSV, PCA, HIS transform, can not meet the demands of all kinds of situations occurring in different fields. This paper presents a new method of Multi-spectral remote sensing images information fusion in order to provide more choices for people. The new method based on k-th central moment can be used to reconstruct a series of images based on the difference between feature vectors and mean vectors of all bands. Comparing with grey image of every band, the reconstructing images´ quantity of information improved greatly as k value is low. With k increased, the reconstructing images´ quantity of information decreased sharply and the difference between feature vectors and mean vectors was amplified. This method can be as a tool to detect the piece of special regions which are different from around environment, like forest fire, environmental pollution, insect pest, flood monitoring, etc.