首页 | 本学科首页   官方微博 | 高级检索  
     

一种基于视觉特性的遥感图像分割
引用本文:王兆虎,刘芳,焦李成.一种基于视觉特性的遥感图像分割[J].计算机学报,2005,28(10):1686-1691.
作者姓名:王兆虎  刘芳  焦李成
作者单位:[1]西安电子科技大学计算机学院,西安710071 [2]西安电子科技大学智能信息处理研究所,西安710071
基金项目:本课题得到国家自然科学基金(60073053,60372045)、国家“八六三”高技术研究发展计划项目基金(2002AA135080)和国家“九七三”重点基础研究发展规划项目基金(2001CB309403)资助.
摘    要:SAR图像存在乘性相干斑的影响,传统的图像分割方法不能很好保持图像的结构信息.文章研究了基于视觉特性的组合分割方法,该方法通过塔型方向滤波器组进行去噪,发展了各向异性的核期望转移滤波器,核的形状、尺度和方向都能自适应图像的局部结构,最后结合边缘检测对图像分割进行修正.分析和实验表明该方法对SAR图像是有效的.

关 键 词:SAR图像分割  塔型方向滤波器组  各向异性的核期望转移滤波器  边缘检测
收稿时间:2004-07-05
修稿时间:2004-07-052005-04-20

A Remote Sensing Image Segmentation Algorithm Based on Vision Information
WANG Zhao-Hu, LIU Fang, JIAO Li-Cheng.A Remote Sensing Image Segmentation Algorithm Based on Vision Information[J].Chinese Journal of Computers,2005,28(10):1686-1691.
Authors:WANG Zhao-Hu  LIU Fang  JIAO Li-Cheng
Abstract:The multiplicative nature of the speckle noise in SAR images is a big problem in the image segmentation. The traditional algorithm (can't) obtain a better performance in keeping structural information of the image. A compounding segmentation algorithm based on vision information is proposed in this paper. The speckle noise could be denoised by pyramid directional filter banks in the image pretreatment. The authors also develop an anisotropic kernel mean shift in which the kernel's shape, scale, and orientation can adapt to the local structure of the image; and at last the authors revise the image segmentation performance via edge detection. The analysis and experimental results on SAR image demonstrate the effectiveness of this algorithm.
Keywords:SAR image segmentation  pyramid directional filter banks  anisotropic kernel mean shift filter  edge detection
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号