基于提取稳定特征的遥感影像变化检测优化

Remote Sensing Image Change Detection Optimization Based on Extracting Stable Features

为了达到减少误检图斑的目的,提出了一种基于提取稳定特征的遥感影像变化检测优化的方法。首先通过多尺度融合进行遥感变化检测,然后通过SURF算法对两次影像的检测结果进行融合提取稳定特征点,最后将两个检测结果进行重叠剔除稳定特征点,从而提高遥感影像变化结果的精确度。实验结果证明了提取稳定特征点的可行性,且表明稳定特征点的引入可以降低遥感影像变化检测结果的误检率,提高检测的正确率,并优化遥感影像变化检测结果。

In order to reduce the false spot detection,a method based on extracting stable features for remote sensing image change detection is proposed.Multi-scale fusion was used to detect the change of remote sensing image,then the stable feature points were extracted from the two images by SURF algorithm.Finally,the two detection results were overlapped to eliminate the image spots where the stable feature points were located,so as to improve the accuracy of the change results of remote sensing image.The experimental results proved the feasibility of extracting stable feature points,and showed that the introduction of stable feature points can reduce the false detection rate of remote sensing image change detection results,improve the detection accuracy,and optimize the remote sensing image change detection results.