Study on mixed pixel classification method of remote sensing image based on fuzzy theory

Classification of mixed pixel is the difficult point of remote sensing image. Fuzzy theory is the effective method of classification of mixed pixel. It can get kind result in classification of mixed pixel. The fuzzy theory is adopted in the classification of remote sensing images, in which each pixel is not only sorted to a single type, but also several feature types. Since one pixel is correlated to several feature type, the degree of correlation to each type is expressed as [0,1], and then the membership degree of the pixel to every features is to be calculated for classification. This paper has carried on systematic theory analysis and experiment to TM remote sensing image with the classification model of the fuzzy theory. A experiment has carried through at aiming to the TM remote sensing images of Fuxin City in 2006. A pre-processing to the remote sensing images is done firstly, by means of MNF (minimum noise fraction transformation) to separate the noise and the signal. Next, the pure pixels are pickup through the pixel purity index images using the end-member selection of pixel purity index (PPI) and the final end-member is determined. At last, the classifying to the mixed pixels is carry trough by using fuzzy classification model, so that the classification result is obtained, and the precision is to be analyzed.