基于结构相似性与直线提取的建筑物变化检测方法

摘要

In this paper, a building change detection method with high accuracy and efficiency is proposed. It is based on LSD line ex-traction algorithm and structural similarity, and high resolution remote sensing images of two different times and vector information of building in DLG were utilized. The detailed steps of the change detection were presented,we conducted experiment to validate the algo-rithm using test data, we also analyzed the influence of threshold to the result.%利用数字线划图中建筑物的矢量信息,根据前后两个时期的高分辨率遥感影像,通过结合LSD直线段提取算法与结构相似性度量,采用比较特征差异的方式提出了一种能够以较高的正确率进行快速建筑物变化检测的算法.本文对检测算法的过程进行了介绍,通过实验数据对算法进行了检验,探讨了阈值变化对结果的影响.