Multispectral Image Classification Based On Texture Analysis

In order to restrain the effect of different objects having the same spectrum and the same objects having different spec-trum,a new algorithm,based on hierarchical analysis of texture and spectrum feature,is present in this paper. Through LBP texture feature,it dynamically clusters to form regions having identical or similar landform,so that the effect to classification of topographical objects,caused by the complexity of each region,can be reduced. Then,according to the texture style of each region,it can be classified again by the corresponding spectrum information,and we gain the finally result of remote sensing image. Taking Daqing re-mote sensing image for example,we give the result of experiment and the evaluation of the accuracy.