基于相似系数的1种算法在遥感图像分类中的应用

An Algorithm for Supervised Classification of Remote Sensing Images

用基于向量夹角余弦的相似系数方法对黄河三角洲 1 998年 1 2月份的 Landsat- 5TM遥感影像进行了分类 ,分类精度分析表明用夹角余弦作为分类的相似性度量指标是合理的。这种遥感图像的分类方法简单 ,在实际分类中可以作为常用的方法之一。

In this paper a method of similarity coefficien ts based on cosine of vectors to classify remote sensing images is presented and analyzed. This method is applied to the classification of the remote sensing ima ges of Landsat 5 TM in the Yellow River Delta, and to some extent the results a re found to be relatively satisfying. Therefore, the method can be used as one o f the routine methods for the classification of the remote sensing images.