基于融合和IFLICM算法的非监督遥感影像变化检测

提出了一种基于多尺度小波融合和改进的非监督模糊聚类的多光谱遥感影像变化检测方法。该算法解决了目前很多算法造成虚警率较高，而且未能充分利用像元之间空间关系的问题。首先利用二维离散小波（DWT）多尺度分解的方式来构造差异图，通过对两种小波分解系数融合的方式来抑制噪声点和突出变化区域。考虑到像元之间的空间位置信息，在融合后的基础上采用改进的模糊局部信息聚类（IFLICM）的方法得到变化检测结果。对两个时相的多光谱遥感卫星影像进行变化检测试验，试验表明基于融合的变化检测结果精度更高，并且改进后的聚类算法效果比其他聚类算法效果更好。

Unsupervised Remote Sensing Image Change Detection Based on Fusion and IFLICM Algorithm

A new method based on multiscale wavelet fusion and improved unsupervised fuzzy clustering algorithm for multispectral remote sensing image change detection is proposed. This algorithm solves the problems that many state of arts algorithms causing too much noise,and failing to fully exploit the spatial relationship between pixels.Firstly,the difference image is constructed by the method of two-dimensional discrete wavelet transform ( DWT) multi-scale decomposition,so the speckle noise is constrained and the changed regions are highlighted by fusing two wavelet decomposition coefficients. Next, Considering the spatial neighborhood information of pixels,the improved fuzzy local information clustering ( IFLICM) method is implemented to get the result of the change detection on the basis of fusion image.Experiments on multi-temporal images show that the image fusion strategy integrates the advantages of CVA and ADI images and gains a better performance.The change detection results obtained by the improved fuzzy clustering algorithm exhibited lower error than other clustering algorithms.