多特征融合的高分辨率影像建筑物变化检测

为充分发挥遥感影像中各特征的优势，提高遥感影像建筑物变化检测精度，基于面向对象的分析方法，提出了一种基于模糊集合的证据理论特征信息融合的变化检测方法。首先，在影像分割的基础上，利用变化矢量分析法分别计算前后时相对应对象的光谱、纹理特征差异及形态学建筑物指数差异；然后，以Sigmoid函数作为隶属度函数，计算对象属于变化类和非变化类的隶属度并以之构建证据理论所需的基本概率分配函数；最后，利用证据理论对多种特征进行融合并通过规则判定得到建筑物变化区域。利用不同地区影像的试验结果表明，该方法能够有效融合影像的多种特征，提高建筑物变化检测的精度。

Building change detection by multi-feature fusion from high resolution remote sensing images

In order to make full of spectral, spatial and texture feature, a method of change detection based on fuzzy set theory and DS evidence theory is proposed. In the first step of the proposed approach,image segmentation is used to get image objects.The spectral, texture and morphological building index features of each object are extracted. Change vector analysis is adopted to calculate the difference of the corresponding features between two periods. Then, Sigmoid function is applied as membership function to get the objects membership belonging to the changing and non-changing classes respectively and to construct the basic probability assignment function(BPAF) of evidence theory.At last, evidence theory is used to fuse multi-feature information and the changed image region of building is determined via certain rules. The experimental results show that the method can fully integrate spectral, texture and morphological building index(MBI) features and improve the accuracy of building change detection.