结合变化向量分析和直觉模糊聚类的遥感影像变化检测方法

针对多时相遥感影像变化检测存在数据不确定性、检测精度不高等问题，提出了一种结合变化向量分析（CVA）和直觉模糊C均值聚类算法（IFCM）的多时相遥感影像变化检测方法. 首先通过CVA构建两个时相遥感影像的差异影像；然后采用直觉模糊C均值聚类算法对差异影像进行聚类得出变化区域和未变化区域；最后对变化检测结果进行二值化处理并进行精度评价. 选取两个时相的高分一号遥感影像和Szada数据集影像作为实验数据. 实验结果表明，采用提出的方法可有效解决传统方法存在的数据不确定性问题，变化检测精度达到了95.92%和92.70%，是一种可行的遥感影像变化检测方法. 研究结果可用于森林动态变化监测、土地复垦利用规划变化分析以及灾损评估. 

Change detection in remote sensing images combined with intuitionistic fuzzy clustering and change vector analysis

Aiming at the problems of multi-temporal remote sensing images change detection with data uncertainty and low detection accuracy, a multi-temporal remote sensing images change detection method combined with change vector analysis (CVA) and intuitionistic fuzzy C-means clustering algorithm (IFCM) is proposed. Firstly, the difference image of bi-temporal remote sensing images is obtained by change vector analysis method. Then the difference image is clustered by the intuitionistic fuzzy C-means clustering algorithm to obtain the change areas and the non-change areas. Finally, the change detection results are binarized and the accuracy is evaluated. The bi-temporal Gaofeng-1 remote sensing images and Szada image data sets were selected as experimental data. The experimental results show that the proposed method can effectively solve the data uncertainty problem existing in the traditional method, it is a feasible remote sensing images change detection method. The overall accuracy of change detection achieved 95.92% and 92.70%. The research results can be used for forest dynamic change monitoring, land reclamation utilization planning change analysis and damage assessment.