基于特征提取的SAR图像滤波算法  

A Filtering Algorithm of SAR Image Based on Feature Extraction

作　　者：王瑞霞[1] 林伟[1] 梅珍[1] 

机构地区：[1]中国西安西北工业大学理学院,西安710072

出　　处：《传感技术学报》2009年第6期827-832,共6页Chinese Journal of Sensors and Actuators

基　　金：国家自然科学基金资助(60375003);航空基础科学基金资助(03I53059)

摘　　要：针对SAR图像所包含的地物目标比较复杂的特点,提出一种新的SAR图像滤波算法。该算法利用Contourlet变换能保持边界,结合特征提取可以获得区域目标点特征的特点,利用自适应阈值的方法对SAR图像进行滤波。与传统Lee滤波和保持边缘特征滤波算法进行比较,实验表明新的算法不仅提高了图像的质量,其量化指标也得到了提高。作为该算法的应用,利用变换域内低频图像和子带图像的融合规则对L和C两个波段的SAR图像进行去噪融合,可以看出该算法的去噪效果。By reason of complexity of objects of SAR image, it puts forward a new filtering method. This method utilizes Contourlet transformation togethering with feature extraction,which can get point feature of areas. Using of auto-adapting threshold filters SAR image. The experimental results show that the results using our methods are better than that of using the other methods such as Lee and keeping edge filters. It materialize that not only the good vision of image were gotten but the quality indices were great improvement. As a practical application, SAR images of two wave bands, L and C , were denoised and fused. And it have got preferable result.

关 键 词：CONTOURLET变换 特征提取 融合 滤波 SAR(Synthetie apestuse radar)图像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]