基于分形理论的遥感蚀变异常提取和分级

Extraction and grading of remote sensing alteration anomaly based on the fractal theory

分形理论研究自然界和社会活动中广泛存在的无序(无规则)而有自相似性的系统。在地球系统中,许多地学现象的空间展布具有分形结构。蚀变作为一种地质记录,其分布具有明显的不均匀性。同时蚀变也是一种异常,而分维与异常又往往具有一定的关系。利用分形理论的求和法(即像元亮度-面积模式)进行遥感蚀变异常的提取和分级较为有效,可以弥补传统的门限化方法和利用灰度直方图目视确定突变点方法的不足。

Fractal theory studies the widespread disordered （ruleless） system with self-similarity in the nature and social activities. Within the earth system, the spatial distribution of many geo-scientific phenomena displays ffactal structure. As a kind of geological record, alteration is clearly not homogeneous. Meanwhile alteration also can be regarded as a kind of anomaly, and fractal dimension is usually linked with anomaly. The summation method （namely pixel grayscale and area mode） based on fractal theory is an effective way to extract and grade alteration anomaly from the background. This method can make up the insufficiency of conventional threshold method and identification of catastrophe point using grayscale histogram.