基于蜂群优化模糊聚类的遥感图像变化检测

针对模糊聚类算法容易陷入局部最优,结合人工蜂群算法的全局最优性,提出一种基于蜂群优化模糊C均值聚类的新算法,并将此算法应用到遥感图像的变化检测中。利用差值图和比值图融合的方法得出多时相遥感图像的差异图,在对差异图像进行模糊聚类生成变化类和未变化类的同时,利用人工蜂群算法对差异图进行全局搜索,较大程度地避免FCM算法陷入局部最优,也降低了FCM算法对初始解的敏感度。实验结果表明,新算法比FCM分类准确、效率更高。

Change Detection in Remote Sensing Images Based on the Fuzzy Clustering Algorithm and Artificial Bee Colony Optimization

In order to overcome the local optimization of the fuzzy clustering algorithm,an artificial bee colony based on fuzzy algorithm combined with the global optimization of the bee colony algorithm is proposed for change detection in remote sensing.Ratio figure and difference figure fusion method is chosen to generate the difference image （DI）,and then the fuzzy clustering algorithm is adopted to recover the changed and unchanged regions of the DI by constructing two clusters,where the artificial bee colony algorithm is introduced to avoid the local minimum problems of FCM and reduce the sensitivity of the initialization values of FCM.Simulation results show the new algorithm is more robust and efficient.