Study About The Recognition Method Of Land Cover With Remote Sensing Image Based On Fractal Texture

With the development of technique about space information,RS and computer network at very fast speed,the obersevation technique toward the earth's surface has also been improved. The recognition and translation of land cover with remote sensing image,as an important branch of geoscience field,has become a hot point of study at home and abroad currently. As important information and basic character,texture has significant meaning in translation of remote sensing image. This thesis study around land cover's texture feature of RS image,puts forward diverse mathematics models of texture classification,via analyzing their advantages and disadvantages,and then brings forward a comparatively reasonable and effective recognition method of remote sensing image based on fractal model of texture.In the first place,this paper studies the definition and their correlative theory foundation of texture and fractal by the numbers. And then setting an example to CBERS-1 image of Three Gorges area, it sets forth and establishes detailed mathematics model of image classification based upon fractal texture. Taking part area Three Gorges as experiment area,according to established model,this paper adopts ameliorated Fuzzy C-Means Clustering idea to achieve classification of land cover. In the end,the author assesses the precision of classification results with Confusion-Matrix,the precision is improved and the ruselt is satisfied. the experiment verifies that the classification method based on fractal texture possesses practical application value and significant theoretical meaning.