运动模糊图像复原技术的算法研究

发布时间：2018-07-17 03:20

【摘要】：随着计算机技术的飞速发展和图像应用的日常化,图像复原成为一个应用面极为广泛的多学科交叉渗透的领域,数字图像领域的研究重点也逐渐转移到对图像复原技术的研究上来。在图像复原领域中,运动模糊图像属于其中较难的一类,它在各个领域的应用都非常普遍。对它的研究具有极其重大的实际意义,也引起了国内外专家的注意,并取得了斐然的成绩。然而,部分科研工作者提出的复原算法的计算相对繁琐复杂,不能很好兼顾图像细节的保存和伪像的消除。本文在广泛的理解和分析前人科研成果的基础上,综合各种方法的优势,提出一种改进的Lucy-Richardson复原算法。首先,描述了国内外在运动模糊图像恢复领域的研究情况,并介绍了相关的图像复原技术的理论基础。对运动模糊图像的退化机理和经典复原算法的工作原理做了叙述,并对各算法分别从主观、客观方面做出了评价,分析它们的优缺点。其次,对于维纳滤波法中不能准确估计K值的问题,提出一种改进的自动估计K值的方法,通过仿真验证改进的K值估计算法可以准确估计K值,提高了复原图像的质量。最后,本文重点分析了Lucy-Richardson算法,通过仿真实验,发现在使用Lucy-Richardson复原算法时暴露出一个很大的缺陷,迭代次数的增多会导致噪声对图像的污染越来越严重,而且迭代次数基本上都是不明确的,这些问题导致了Lucy-Richardson算法的应用范围受到限制。针对这些问题,提出的改进算法先是对模糊图像用高斯滤波器进行去噪声处理以抑制振铃效应,并在迭代公式中加入系数项用来约束复原图像。对迭代次数不能准确估计的问题,提出合理的解决办法,即最佳迭代次数的范围由大步长来确定,而精确值由小步长来确定。最后利用Matlab软件来仿真改进的Lucy-Richardson算法的效果。对仿真实验的结果从主观和客观两个方面评价,由评价参数可以证明本文算法更加有效。
[Abstract]:With the rapid development of computer technology and the daily application of images, image restoration has become a wide range of multidisciplinary cross-penetration field. The research focus of digital image field has gradually shifted to the research of image restoration technology. In the field of image restoration, motion blur image belongs to one of the more difficult one, and its application in every field is very common. The research on it is of great practical significance and has attracted the attention of experts at home and abroad, and has made remarkable achievements. However, the computation of the restoration algorithm proposed by some researchers is relatively complicated, which can not take into account the preservation of image details and the elimination of artifacts. On the basis of extensive understanding and analysis of previous scientific research achievements, this paper proposes an improved Lucy-Richardson restoration algorithm by synthesizing the advantages of various methods. Firstly, the research situation in the field of motion blur image restoration at home and abroad is described, and the theoretical basis of the related image restoration technology is introduced. The degradation mechanism of motion blur image and the working principle of classical restoration algorithm are described. The subjective and objective aspects of each algorithm are evaluated and their advantages and disadvantages are analyzed. Secondly, for the problem that the K-value can not be estimated accurately by Wiener filter, an improved method of automatically estimating K-value is proposed. The simulation results show that the improved K-value estimation algorithm can estimate the K-value accurately and improve the quality of restored image. Finally, the Lucy-Richardson algorithm is analyzed in detail. The simulation results show that the Lucy-Richardson algorithm has a big defect when it is used. The increase of iteration times will lead to more and more noise pollution to the image. Moreover, the number of iterations is not clear, which leads to the limitation of the application of Lucy-Richardson algorithm. To solve these problems, the improved algorithm is first used to remove noise from fuzzy images using Gao Si filter to suppress the ringing effect, and the coefficient term is added to the iterative formula to constrain the restoration of the image. For the problem that the number of iterations can not be estimated accurately, a reasonable solution is put forward, that is, the range of the best iteration number is determined by the large step size, and the exact value is determined by the small step length. Finally, the improved Lucy-Richardson algorithm is simulated by Matlab software. The results of simulation experiments are evaluated from both subjective and objective aspects. The evaluation parameters can prove that the proposed algorithm is more effective.
【学位授予单位】：北京印刷学院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【相似文献】