面向对象的遥感影像变化检测方法

针对变化检测区域内变化区域与未变化区域面积比例较低时,通过常规的阈值计算无法在变化检测中确定准确的变化阈值问题,该文提出了一种带样本选择的面向对象遥感影像变化检测方法。该方法首先对多时相遥感影像进行多尺度分割获取像斑,并采用变化向量分析法计算像斑的差异度;然后,自适应选择训练样本,结合基于期望最大化算法和贝叶斯最小误差率理论的阈值计算方法,采用独立阈值法确定变化阈值;最后,利用变化阈值对差异影像进行二值分割,并获取变化检测结果。实验结果表明该文方法在变化检测精度上优于常规方法。

Object-based change detection method of remote sensing images

The accurate change threshold could not be worked out by the general global or local threshold method,if the prior probability of the class of changed pixels in the detection region is low.In order tosolve this problem,an object-based change detection method of remote sensing images based on sample selection was proposed.Firstly,multi-scale image segmentation was used to get image objects from the multi-temporal remote sensing images,and difference of image objects was calculated from each image object based on change vector analysis (CVA).Then,the training samples were automatically selected.Based on the expectation maximization (EM) and Bayesian rule with minimum error rate,the change threshold could be worked out by the independent threshold method.Finally,difference image was binary segmented by the change threshold,and the change detection result was derived.The results showed that the proposed method was superior to general methods in the change detection accuracy.