针对不同时相遥感影像变化检测研究中变化区域的自动提取问题,本文提出一种基于类间最大方差准则,利用C均值算法自动确定变化阈值的方法。该方法中将变化区域提取问题转化为两类之间的分类问题,利用C均值算法进行迭代处理,当两类之间方差最大时即为最佳变化阈值T。实验结果表明,该方法可准确快速地确定图像变化检测的最佳阈值,实现变化区域提取的自动化。

Extracting changed areas from different images was an important problem in the field of remote sensing image change detection.To solve this problem,a method based on maximal variance between-class criteria and C-means algorithm was proposed.Changed area extraction was converted into a typical problem of two-category classification and could be solved by employing threshold strategy.The C-means algorithm is used to classify an image into two classes and obtained its best threshold when the variance between-class is maximal.The experimental results show that the method can automatically determine the best image change detection threshold and extract the changed areas quickly and accurately.