| 基于SVM的SAR图像分割方法 |
| |
| 引用本文： | 靳红梅,刘蓉,张俊梅,梁荣.基于SVM的SAR图像分割方法[J].北京服装学院学报,2009,29(3). |
| |
| 作者姓名： | 靳红梅 刘蓉 张俊梅 梁荣 |
| |
| 作者单位： | 西安科技大学计算机系,西安,710054;北京服装学院基础教学部,北京,100029 |
| |
| 摘 要： | 先去噪再分割的SAR图像分割方法会损失有用的纹理信息,因此本文给出了一种直接对含噪SAR图像进行分割的方法.首先利用小波变换提取SAR图像的纹理特征,计算图像的灰度均值作为图像的灰度特征,然后用完全无监督的聚类算法进行分类,最后将特征值与类别标记作为支持向量机的训练样本,用训练后的分类器对图像进行分割.实验结果表明:本文所给出的方法在分割的准确性和抗噪性方面都优于几种有代表性的分割方法.
|
| 关 键 词： | SAR图像分割 小波变换 SVM |
SAR Image Segmentation Method Based on SVM |
| |
| JIN Hong-mei,LIU Rong,ZHANG Jun-mei,LIANG Rong.SAR Image Segmentation Method Based on SVM[J].Journal of Beijing Institute of Clothing Technology(Nature Science Edition),2009,29(3). |
| |
| Authors: | JIN Hong-mei LIU Rong ZHANG Jun-mei LIANG Rong |
| |
| Affiliation: | JIN Hong-mei1,LIU Rong2,ZHANG Jun-mei1,LIANG Rong1 (1.Department of Computer,Xi\'an University of Science , Technology,Xi\'an 710054,China,2.Department of Fundamental Courses,Beijing Institute of Fashion Technology,Beijing 100029,China) |
| |
| Abstract: | The classical SAR image segmentation was performed through speckle filter,so some useful texture information disappeared.The method of SAR image segmentation based on SVM was presented in this paper.At first,the texture features were distilled with wavelet,the mean of the regional gray-level was used to extract gray-level features,and then the unsupervised clustering algorithm was used for classification.The energy of region and labels were taken as training data of SVM to support vector machine.In the end,... |
| |
| Keywords: | SVM |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
