基于自适应小波阈值与曲波变换的SAR图像去噪  

SAR Image Denoising Based on Adaptive Wavelet Threshold and Curvelet Transform

作　　者：杨哲 邵哲平[1,2] YANG Zhe;SHAO ZhePin(Navigation Institute,Jimei University,Xiamen 361021,China;National-Local Joint Engineering Research Center for Aid to Navigation, Jimei University, Xiamen 361021, China)

机构地区：[1]集美大学航海学院,福建厦门361021 [2]集美大学船舶辅助导航技术国家地方联合工程研究中心,福建厦门361021

出　　处：《中国航海》2020年第4期46-51,共6页Navigation of China

基　　金：国家自然科学基金(51879119);福建省自然科学基金(2018J05085)。

摘　　要：目前,国内外都使用合成孔径雷达(Synthetic Aperture Radar,SAR)做为海上小目标检测应用的重要技术手段,然而海上小目标的检测存在SAR图像不够清晰的问题。针对该问题,设计将自适应小波阈值与曲波变换去噪方法相结合,对图像进行去噪处理。该融合算法不仅能够解决传统小波阈值去噪“一刀切”造成的过度去噪以及只能对“点奇异”特征函数的最优逼近,而且优化曲波变换只对“线奇异”特征函数最优逼近的问题。通过一系列的量化指标和视觉效果进行去噪性能比较可知:该方法背景杂波更加平滑、目标与背景对比度更高。SAR(Synthetic Aperture Radar)is used as an important technical means for detecting small targets at sea generally.But all small targets at sea can not be clearly distinguished in SAR images.The adaptive wavelet threshold and curvelet transform are integrated to carry out image denoise,making the targets more distinguishable.The integration of the two processes can solve the problems that would happen otherwise if only one of them were used.The fixed threshold would over denoising and the wavelet threshold denoising algorithm can only approximate the“point singularity”eigenfunction while the curvelet can only approximate the"line singularity"eigenfunction.The effectiveness of the algorithm is demonstrated by the denoising comparison of quantitative indicators and visual effects,knowing that background clutter is smoother and the contrast between background and target is better.

关 键 词：SAR图像去噪 自适应小波阈值 曲波变换 

分 类 号：U675.74[交通运输工程—船舶及航道工程]