不同类型噪声的四阶偏微分方程滤波方法的研究及其应用-光学专业论文.docxVIP

摘 要 图像信号在其形成、传输、变换、以及终端的处理过程中，经常会受到各种 噪声的干扰而降质，主要包括：散斑噪声、脉冲噪声、高斯噪声和泊松噪声等。 图像滤波是图像预处理的一个重要的环节，其处理效果的好坏将直接影响到后续 图像处理和分析的有效性和可靠性。本论文主要针对于电子散斑干涉（ESPI）条 纹图、相位图中的散斑噪声和标准图像中的脉冲噪声的滤波问题展开研究，提出 了新的四阶偏微分方程、方向四阶偏微分方程以及基于检测技术的四阶偏微分方 程的滤波方法。 ESPI 是一种无损的全场光学测量技术，被广泛应用于无损检测、光学粗糙 表面位移、变形测量、振动分析等领域。然而，散斑在成为信息载体的同时也不 可避免地成为了噪声，ESPI 条纹图和相位图总是伴有强烈的散斑噪声，严重影 响了 ESPI 相位的提取。针对于含有强烈散斑噪声的 ESPI 条纹图和相位图，本文 提出了的适合处理粗条纹的新四阶偏微分方程和适合处理细密条纹的方向四阶 偏微分方程，并与相关的偏微分方程模型的滤波结果进行了比较。实验结果证明 了提出的这两个方程能够在滤除噪声的同时保护边缘。 另外，针对于含有脉冲噪声的标准图像，本文提出了基于检测技术的四阶偏 微分方程。该方法利用了两步复原法的“先检测再滤波”的思想来对图像进行滤 波，即先用 TSM 方法对噪声进行检测，然后用新的四阶偏微分方程对噪声点进 行滤波。实验结果证明了基于检测技术的四阶偏微分方程滤波模型能够移除噪声 并且很好的保护图像的细节。 关键词： 图像滤波； 散斑噪声； 脉冲噪声； 电子散斑干涉； 标准图像； 四 阶偏微分方程 ABSTRACT Image signals are often suffered from noise which including Speckle noise, Impulsive noise, Gaussian noise and Poisson noise in acquisition, transmission, conversion or terminal processing, which usually degrades the visual quality of the images. Image filtering has become an important part of image preprocessing, whose result will have an important and direct effect on the validity and reliability of the image analysis and image recognition. In this paper, the research is mainly focused on the filtering of the speckle noise of the electronic speckle pattern interference technology (ESPI) fringe patterns and phase patterns and the impulse noise of the standard image, and it proposed the new fourth-order partial differential equation (PDE), the fourth-order oriented PDE and the fourth-order PDE with the detection techniques filtering model. ESPI is a kind of non-destructive, whole-field optical measurement technique, which has been widely used in nondestructive testing, optical rough surface displacement, deformation measurement, vibration analysis, and other fields. However, Speckle become the information carrier while becoming the noise, and the ESPI fringe patterns and phase patterns always have strong noises which seriously affect the ESPI phase extraction. For the ESPI fringe patterns