高分辨率图像复原正则稀疏表示方法.pdfVIP

高分辨率图像复原的正则稀疏表示方法 黄石生、王志强、庄家礼、刘安娜 北京跟踪与通信技术研究所 摘要：本文提出一种基于正则稀疏表示的高分辨率图像复原方法。给定观测的模糊图像，传统的稀疏表示复原 方法基于图像在表示基上的稀疏性和直接观测数据建立l 1 范数正则模型。由于模糊算子的紧致性，该模型具有较大 的相关性。本文设计一个正则算子作用于模糊图像，并且基于转化后的数据建立l 1 范数正则模型。我们称其为正则 稀疏表示方法，具有如下优点：1）降低l 1 范数正则模型的相关性；2 ）基于预条件策略，加速模型求解算法的收敛 速度。数值实验结果表明该方法相比于一些已有方法在速度和复原性能上具有更优的效果。 关键词：正则化；稀疏表示；高分辨率；图像复原 中图分类号 TP391.41 文献标识码 A Regularized Sparse Representation And Its Application In High-resolution Image Deconvolution Abstract: A high-resolution image deconvolution approach based on regularized sparse representation is proposed. Given an observed blurred image, traditional deconvolution approach based on sparse representation constructs the l 1 -norm regularization directly by using the observed model and the sparsity of image decomposed on a redundant dictionary. However, the l 1 -norm regularization has large coherence because of the compactness of blurring operator. Therefore, we multiply the blurred image with a regularizing operator and construct the l 1 -norm regularization by using the converted data. The approach, referred as regularized sparse representation, possesses the following merits: i) it decreases the coherence of the l 1 -norm regularization; ii) algorithm solving the regularization is fast converged due to the preconditioning strategy. Experimental results demonstrate that comparing with existed state-of-the-art approaches the proposed approach can obtain good results in terms of speed and deconvolution performance. Keyworkds: regularization; sparse representation; high-resolution; image deconvolution 一、概述 图像是人类获取信息的主要载体。光学传感器在获取图像时，常受到大气传输、平台振动等因 素影响，使得获取的图像降质变得模糊。图像模糊是光学成像系统固有的，对于高分辨率遥感成像 系统，图像模糊问题不容忽视。由于分辨率与像元尺寸成正比、与焦距成反比，对于高分辨率成像 系统，高分辨率使得系统焦距加长、像元减小，因此大气、振动等因素对成像质量的影响尤为严重。 此时，可通过复原处理由模糊图像