利用Brushlet变换进行SAR图像变化检测  

SAR image change detection based on Brushlet transform

作　　者：颜学颖[1] 焦李成[1] 王凌霞[1] 

机构地区：[1]西安电子科技大学智能感知与图像理解教育部重点实验室,陕西西安710071

出　　处：《西安电子科技大学学报》2013年第6期67-73,共7页Journal of Xidian University

基　　金：国家自然科学基金资助项目(61173092;61072106;60971128;61077009;60972148;60970066;61003198;61001206;61050110144);高等学校学科创新引智计划(111计划)资助项目(B07048);教育部"长江学者和创新团队发展计划"资助项目(IRT1170)

摘　　要：针对传统空域和小波域检测算法的相邻像素间相似特征捕捉性能差、方向分辨率低的问题,提出了一种基于非下采样Brushlet变换和各向异性Gabor窗的二维最大类间方差变化检测方法.将非下采样Brushlet域的各向异性Gabor非线性加权均值计算和空域最小化均方误差的线性组合相结合,来获取相干斑噪声抑制后的均值特征,解决了角分辨率低的问题,获得了各个方向、频率和位置的精确定位;利用二维最大类间方差阈值分割来得到最终的变化检测结果.对真实的SAR图像进行了实验,证明了新方法有着较好的检测结果,并能够很好地保留边缘等细节信息.The traditional change detection method has a poor accuracy for similarity character capture and low direction-resolution. In this paper, a new 2D-Otsu SAR image change detection method is proposed based on the overcomplete Brushlet transform and Gabor window. This method combines the local anisotropic Gabor weighted nonlinear mean procedure in the overcomplete Brush[et domain and linear combination with the minimum mean squared error in the original domain to obtain mean character after the speckle noise is removed, which resolves the problem of low direction-resolution, and can accurately position the texture of each direction, frequency and position. Finally, change detection is processed by the 2D-Otsu method which combines the mean character and gray-level character. Experiment results show that the new method has a better performance, and can well preserve the detailed information such as the texture and edge.

关 键 词：图像变化检测 BRUSHLET变换 各向异性 阈值分割 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]