【作者】 王悦；

【导师】 刘婧；

【作者基本信息】 大连海事大学 ， 应用数学， 2011， 硕士

【摘要】 人耳识别技术是新兴的生物特征识别技术.人耳图像在采集过程中会因采集器产生的噪声、采集状态的不同而产生位移、旋转和形变,导致图像产生伪特征信息.因此,在提取人耳特征前必须对人耳图像进行滤除噪声、增强图像特征等图像平滑工作.可以说平滑效果的好坏直接决定人耳识别系统的有效性和准确率.常见的传统平滑滤波,如高斯滤波、均值滤波等在滤除噪声的同时将图像边缘的一些重要信息也一并滤除了.近年来,应用偏微分方程(Partial Differential Equations,PDE)处理数字图像的方法引起人们的广泛关注,这种方法在图像滤波、分割、复原等众多领域都取得了丰硕的研究成果.本文的结构如下：第一章绪论部分概述了论文选题的目的和意义,介绍了图像噪声、传统的图像滤波方法、用偏微分方程进行图像处理、图像处理评价指标等内容,最后是本文的主要研究内容和组织结构安排.第二章介绍了在图像平滑中偏微分方程方法的发展过程和作用,对其中几个平滑模型进行了剖析.第三章是本文的主要工作,具体建立了两个改进的非线性扩散模型并用于处理人耳图像,对模型进行数值计算,选取合适的参数,运用Matlab进行仿真实验并分析结果,总结出改进模型的优点.更多还原

【Abstract】 Ear recognition technology is emerging biometric technology. It would generate pseudo feature information for the noises of collector and different collection state produce displacement, rotation and deformation characteristics of the image on gathering ear image. Hence, before extracting the features of the human ear, we must filter noise, enhance image etc, the quality of smooth effect directly decides the ear recognition system effectiveness and accuracy. Traditional filter methods, such as Gaussian smoothing filter, mean filter and so on, inherently could not respect local features. The price paid for the removal of noise is the decrease of spatial resolution, which is caused by the flattening of sharp edges. In recent years, base on partial differential equations (Partial Differential Equations, PDEs) smoothing methods of image received widely attention in the field of digital image processing. And in many classic images processing issues, such as image filtering, restoration, segmentation, enhancement and other fields have achieved fruitful results.This paper is structured as follows:The first chapter is the introduction, an overview of the research topics of the purpose and meaning, introducing several conventional filtering methods, image processing evaluation criterion, later given the research work contents of this paper, and organizational structure.The second chapter is Partial Differential Equations in the development of image smoothing, and several typical PDE smoothing models are analyzed.The main attention is put on the chapter three, specifically the establishment of two improved nonlinear diffusion model for the human ear image processing, and their numerical solutions of the models, select the parameters for the use of Matlab simulation. Finally, the results summarized the advantages of the improved model.更多还原