| 基于改进自调整图谱方法的SAR图像分割 |
| |
| 引用本文： | 公徐路,田铮,赵伟.基于改进自调整图谱方法的SAR图像分割[J].电子设计工程,2014(5):127-130. |
| |
| 作者姓名： | 公徐路 田铮 赵伟 |
| |
| 作者单位： | [1]西北工业大学理学院,陕西西安710129 [2]西北工业大学计算机学院,陕西西安710129 [3]中国科学院遥感应用研究所遥感科学国家重点实验室,北京100101 |
| |
| 基金项目： | 国家自然科学基金（60972150;61201323;10926197） |
| |
| 摘 要： | 谱聚类是在给定数据集上用基于图论的方法进行分类,并已广泛地应用于SAR图像分割.自调整谱聚类（self-tuning spectral clustering,简称STSC）方法是一种可以自动确定尺度因子和分类数的方法.本文给出了一种改进的STSC方法,使用熵函数作为自动求分类数的代价函数,使得分类数的计算更加准确和有效,提高了方法的分类精度.实验表明,改进的STSC方法对自然图像、SAR图像的分割精度高于原STSC方法.
|
| 关 键 词： | 合成孔径雷达 自调整图谱方法 分割 代价函数 |
SAR image segmentation based on improved self-tuning spectral clustering method |
| |
| GONG Xu-lu,TIAN Zheng,ZHAO Wei.SAR image segmentation based on improved self-tuning spectral clustering method[J].Electronic Design Engineering,2014(5):127-130. |
| |
| Authors: | GONG Xu-lu TIAN Zheng ZHAO Wei |
| |
| Affiliation: | (School of Science, Northwestern Polytechnical University, Xi＇an 710129, China; School of Computer,Northwestern Polytechnical University, Xi＇an 710129, China; State Key Laboratory of Remote Sensing Science, Institute of Remote Sensing Applications, Chinese Academy of Sciences, Beijing 100101, China) |
| |
| Abstract: | Spectral clustering is a classification method on a given data set based on graph theory and has been widely used in SAR image segmentation.Self-tuning spectral clustering method （STSC） can determine the scale parameter and estimate the number of clustering automatically.This paper propose an improved STSC method which uses entropy function as the cost function,making the calculation of clustering number more accurate and effective and improving the clustering precision.The experimental results show that the segmentation performances of optical images and SAR images of the proposed method are better than the standard STSC method. |
| |
| Keywords: | Synthetic Aperture Radar （SAR） self-tuning spectral clustering method segmentation cost function |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
