一种运动模糊复原图像的图像增强算法.docVIP

  一种运动模糊复原图像的图像增强算法# 赵琳* （北京理工大学自动化学院，北京 100086） 5 10 15 摘要：采用 Radon 变换和自相关方法结合超分辨力复原算法对重度模糊图像进行复原可以 取得较好的复原效果，但是模糊图像的噪声在复原过程中会被放大。经研究针对复原图像提 出一种基于最小二乘回归的增强算法，将离散像素分割成连续的若干子区间，对图像的行与 列分别分割成若干子区间以简化像素之间的相关性，采用一条曲线对子区间内的像素进行拟 合并引入权重因子用于融合相邻的两条拟合曲线。针对该算法开展的试验结果显示其在对复 原图像起到明显增强效果的同时亦能有效抑制噪声。 关键词：数字图像处理 混合模糊；联合 PSF；自相关，Radon 变换，超分辨力 中图分类号：TN911.73 Image Enhancement of Restored Motion Blurred Images Zhao Lin (School of Automation, Beijing Institute of Technology, Beijing 100086, China) Abstract: Motion blurred images could be restored by Super-Resolution arithmetic with Radon transformation and autocorrelation even though the blur of images was very severity. Unfortunately the noise of blurred images would be amplified while we captured useful 20 25 30 35 40 information. An enhancement algorithm was proposed in this discourse to improve the quality of the low signal-noise ratio images obtained through restoration arithmetic. The algorithm could eliminate unwanted noises and remain desired signals and was based on the principle of the least square error method, which fitted discrete pixels to continuous piecewise curves. The interval of each row and column was subdivided into several subintervals to simply the fitting of pixels. Then a curve was used to fit the pixels within the subinterval. A weighting technique with a linear weighting factor was proposed to merge two adjacent lines together. A series of experiments were carried out to study the effects of the algorithm. And the signal-noise ratio showed that the proposed algorithm could achieve high quality enhancement images. Key words: digital image processing mixed blur, combined PSF, autocorrelation, Radon translation, Super-resolution 0 引言 如何从模糊图像中尽可能多地提取有用信息一直是公安刑侦、物证鉴定等应用所追求的 重要目标。造成图像模糊的原因有很多种，其中由于在曝光过程中成像设备与目标之间的相 对运动使图像沿某一方向移动而造成的运动模糊是一种典型情况[1][2]。复原此类图像的研究 关键是提出精确估计运动参数算法(运动方向和模糊长度)以及合适的盲复原处理算法。 近年来，Radon 变换被广泛应用于信号处理和数字图像处理领域，它于 1917 年由数学 家 Radon 在积分几何研究中首次提出，其在图像重建中表现为截面函数沿着直线的线积分 [3] 比度，突出图像的细节信息。