首页 | 本学科首页   官方微博 | 高级检索  
     检索      

高分辨率SAR图像分割及目标特征提取
引用本文:高贵,匡纲要,李德仁.高分辨率SAR图像分割及目标特征提取[J].宇航学报,2006,27(2):238-244.
作者姓名:高贵  匡纲要  李德仁
作者单位:1. 国防科技大学电子科学与工程学院一系434房间,长沙,410073
2. 武汉大学,武汉,430079
摘    要:提出了一种非监督SAR图像快速分割算法,对分割得到的目标区域进行特征提取.分割方法利用固体热扩散模型与图像尺度空间的等价性,在SAR图像初始分割的基础上,引入最大后验概率矩阵的各向异性多尺度平滑,在保持图像结构信息的同时滤除斑点噪声对于分割的影响.然后提取目标区域6个特征以详细描述目标,实验结果表明:该方法计算速度快,能够从获得的目标区域得到大量有用的信息.

关 键 词:合成孔径雷达  图像  各向异性  热扩散方程  特征提取
文章编号:1000-1328(2006)02-0238-07
收稿时间:04 21 2004 12:00AM
修稿时间:2004-04-212005-09-09

High-Resolution SAR Image Segmentation and Target's Feature Extraction
GAO Gui,KUANG Gang-yao,LI De-ren.High-Resolution SAR Image Segmentation and Target''''s Feature Extraction[J].Journal of Astronautics,2006,27(2):238-244.
Authors:GAO Gui  KUANG Gang-yao  LI De-ren
Institution:1. Room 434, Department 1, SESE, NUDT, Changsha 410073, China; 2. Wuhan University, Wuhan 430079, China
Abstract:An unsupervised fast segmentation method for SAR image is proposed in this paper, and features are extracted from the target region gained from segmentation. According to the equivalence between the solid heat diffusion model and image scale-space, and on the basis of initial segmentation of SAR image, This method includes the multi-scale anisotropic smoothing of the posterior probability matrixs to get rid of the effect of speckle and preserve the important structure information of image. Then 6 features of target area are extracted to describe target in detail. The experiment results show that this algorithm is fast and can get lots of information from target areas.
Keywords:Synthetic aperture radar  Image  Anisotropic  Heat diffuse equation  Feature extraction
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