【摘要】 遥感图像的一个典型应用是进行地面的变化监测。美国Landsat系列卫星提供的TM图像提供了进行变化监测的良好数据源。文章在对同一地区不同时相的TM图像精配准的基础上,对K-均值聚类差异法的实现过程进行了较为详细的探讨,并进行了多波段KL变换、主分量差异法、差异主分量法等处理,以提取该地区由于工程或其它人类活动所造成的地表破坏在遥感图像上所反映的变化信息,比较了通过这些方法来提取多光谱遥感图像上具有一定范围的变化信息的优缺点。更多还原

【Abstract】 A typical application of remote sensing image is to suspect variety information on the Earth.TM image ac-quired by Landsat serial satellites is a good data source to inspect variety information.This paper firstly registers the TM image acquired on different days of the same area,and then carries through multi-band KL transform,PCA differ-entia,differentia PCA,and K-mean clustering differentia.Especially,the implementation of K-mean clustering arithmetic is discussed in detail.And then This paper detects the variety information on the result image and discusses the advantage and disadvantage of each method.更多还原