基于时空自相关的建筑物变化检测

Building Change Detection Based on Spatio-temporal Autocorrelation

针对传统基于像素级的变化检测方法在变化分析中难以利用像元间的时空关系、变化检测结果精度低的问题,提出了一种基于时空自相关的建筑物变化检测方法。首先,利用形态学建筑指数(Morphological Building Index,MBI)进行建筑物提取,并通过长宽比、面积等剔除道路信息优化建筑物提取;其次,采用时空自相关模型分别构建两期MBI特征影像的时空自相关性指标值作为对应像元的相似性测度;最后,利用最大类间方差(otsu)法确定最优阈值,得到变化检测结果。实验表明,该方法所得变化检测结果更完整,漏检率和误检率均低于对比算法,该方法基本满足变化检测需求,为高分影像建筑物变化检测提供一种新的技术手段。

Aiming at the problem that the traditional pixel-based change detection method is difficult to use the spatial-temporal relationship between pixels and the low accuracy of change detection results in the change analysis of building,a detection method of building change based on Spatial-Temporal Autocorrelation is proposed. Firstly,the Morphological Building Index( MBI) is used to extract buildings,and the road information is eliminated to optimize the building extraction. Secondly,the spatio-temporal autocorrelation model is used to construct the spatio-temporal autocorrelation measure values of two MBI feature images as the similarity of corresponding pixels. Finally,the optimal threshold is determined by the objective function method to obtain the change detection results.The experiment shows that the change detection results obtained by this method are more complete,and the rate of missed detection and false detection are lower than the contrast algorithm. This method basically meets the change detection needs and provides a new technical means for the change detection of construction land.