基于非下采样Contourlet变换的SAR图像增强.pdfVIP

第 31 卷第 7 期 电 子 与 信 息 学 报 Vol.31No.7 2009 年 7 月 Journal of Electronics Information Technology Jul..2009 基于非下采样Contourlet 变换的SAR 图像增强 沙宇恒 刘 芳 焦李成 (西安电子科技大学信息处理研究所和智能感知与图像理解教育部重点实验室 西安 710071) 摘 要：该文基于非下采样 Contourlet 变换(NSCT)和 SAR 图像的统计特性，提出一种 SAR 图像增强方法，给出 一种基于非下采样塔型分解的斑点噪声方差估计算法和一种基于方向邻域模型的弱边缘增强算法。该文在不同方向 子代进行斑点方差估计，利用局部方向统计信息对 NSCT 系数并进行强边缘、弱边缘和噪声分类并进行弱边缘的 增强和噪声的抑制。实验结果表明，该方法在方向信息保留和斑点抑制上优于非下采样小波变换(NSWT)相应方法。 关键词：SAR 图像增强；几何多尺度分析；非下采样 Contourlet 变换；非下采样小波变换 中图分类号：TP751 文献标识码：A 文章编号：1009-5896(2009)07-1716-06 SAR Image Enhancement Based on Nonsubsampled Contourlet Transform Sha Yu-heng Liu Fang Jiao Li-cheng (Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education of China, Institute of Intelligent Information Processing, Xidian University, Xi’an 710071, China) Abstract: Based on nonsubsampled Contourlet transform and SAR image statistical property, a SAR image enhancement method is proposed. A speckle noise variance estimate algorithm is given using nonsubsampled Laplace Pyramid decompose, and a wake edge enhancement algorithm using directional local neighborhood is proposed. This paper estimate the speckle variance in each decomposes direction, and the directional local neighborhood statistical is used to distinguish the strong edge, wake edge and noise. The wake edge is enhanced and the speckled noise is restrained. Experiment results show that the method represents better performance compared with NSWT in wake edges information enhancement and speckle reduction. Key words: SA