第十一讲图像复原详解.pptVIP

Chapter 6 Image Restoration Explain Content 图像恢复的概念、模型与方法Purpose 熟悉线性平移不变系统图像退化模型，掌握逆滤波、维纳滤波恢复方法 一、Examples of Image Degradation 二、 Basic Conception 图像的退化: 是指图像在形成、传输和记录过程中，由于成像系统、传输介质和设备的不完善，使图像的质量变坏。 图像增强和图像复原的差别 图像增强不考虑图像是如何退化的，而是试图采用各种技术来增强图像的视觉效果。因此，图像增强可以不顾增强后的图像是否失真，只要看得舒服就行 而图像复原就完全不同，需知道图像退化的机制和过程等先验知识，据此找出一种相应的逆处理方法，从而得到复原的图像。 如果图像已退化，应先作复原处理，再作增强处理。 二者的目的都是为了改善图像的质量。 Image Restoration vs Image Enhancement Both processes are based to improve image quality. Image enhancement is largely a subjective process, while image restoration is for the most part an objective process. Enhancement techniques are heuristic while restoration techniques are mathematical. 三 The Math Model of Image Degradation Problem Description Input: a degraded image g(x,y), some knowledge about the degradation function H(x,y), some knowledge about the additive noise term η(x,y) Objective: an estimate of f(x,y) 三、退化的数学模型 Noise model Noise sources Acquisition: light levels, sensor temperature Transmission: interference in the channel Assumptions: noise is independent of spatial coordinates (except for periodic noise) and independent of the image Spatial description of noise: statistical behavior of the values of the noise Noises are taken as random variables Random variables, which is characterized by a Probability density function (PDF). Notes to Impulse Noise Impulse noise is also called bipolar impulse noise, if either Pa or Pb is zero, the impulse noise is called unipolar. Bipolar impulse noise is also called salt-and-pepper noise, or shot and spike noise. Noise impulses can be negative or positive. Negative impulses appear as black (pepper) points. Positive impulses appear as white (salt) noise. Restoration in the presence of noise Additive noise only 五 经典复原方法之一——逆滤波 (Inverse Filtering) 逆滤波复原 五 经典复原方法之二——维纳滤波 (Wiener Filtering) Wiener Filtering Wiener Filtering Example of Wiener Filtering 六 估计退化函数（Estimating the Degrad