考虑时空关系的遥感影像变化检测和变化类型识别

Remote Sensing Image Change Detection and Change Type Recognition Based on Spatiotemporal Relationship

摘要: 提出了一种基于时空关系的遥感影像变化检测及类型识别方法。该方法通过影像分割获取像斑，利用最大似然法(maximum likelihood method，ML)获取初始分类结果，通过地物的类别转移矩阵（class transition matrix，CTM）和类别邻接矩阵（class adjacency matrix，CAM）定量地描述各地物类别之间的时间关系和空间关系。实验结果显示，本方法优于分类后比较法。

Abstract: A method for change detection and change type recognition of remote sensing images based on spatiotemporal relationship is proposed in this paper. Image segmentation is adopted to get image segments for original image classification using maximum likelihood method(ML). Quantitative spatiotemporal relationship is described by class transition matrix(CTM) and class adjacency matrix(CAM). In order to verify the validity of the algorithm, the method proposed in the paper is compared with the post classification comparison(PCC) method. Experimental results confirm that the proposed method can provide higher accuracy.