基于稀疏和低秩Hankel矩阵分解的SAR图像相干斑抑制  

SAR Image Despeckling Based on Sparse and Low-rank Hankel Matrix Decomposition

作　　者：蔡金萍 赵曜 CAI Jinping;ZHAO Yao(School of Information Engineering,Guangdong University of Technology,Guangzhou,Guangdong Province,510006 China)

机构地区：[1]广东工业大学信息工程学院,广东广州510006

出　　处：《科技创新导报》2020年第32期69-75,共7页Science and Technology Innovation Herald

基　　金：国家自然科学基金青年项目(项目编号:61907008)。

摘　　要：合成孔径雷达(Synthetic Aperture Radar,简称SAR)是现代遥感技术中的一个研究热点。当雷达波遇到粗糙表面时,由于返回信号中各基本散射体之间的相位不同,导致回波之间发生干涉,从而使得SAR图像中包含大量的相干斑噪声,影响了图像的质量及后续的图像解译。本文提出一种基于稀疏和低秩Hankel矩阵分解的方法,与传统的方法相比,有效地改善图像质量,抑制相干斑噪声,保持图像边缘尖锐的同时使得成像后图像的平滑性能得到了更好的提升。通过模拟实验和真实数据实验,可以得出本文所提算法在SAR图像相干斑抑制方面取得了较好的效果。Synthetic Aperture Radar(SAR)is a research hotspot in modern remote sensing technology.When the radar wave reaches the rough surface,interference between the echoes occurs due to the difference of basics scatters phases.That causes a lot of speckle noise in the SAR image and affects the quality of the SAR image.In this paper,a method based on sparse and low-rank Hankel matrix decomposition proposes.Compared with the traditional algorithm,this method ef fectively improves the quality of the image and despeckles noise.While keeping the edge sharp,the smoothness of images is better improved.Through simulation experiments and real data experiments,it can conclude that the proposed algorithm has achieved better results in the aspect of SAR image speckling.

关 键 词：合成孔径雷达 相干斑抑制 低秩Hankel矩阵分解 稀疏 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]