数字图像处理在指纹识别中的应用 毕业设计.docx

 PAGE \* MERGEFORMAT 45 数字图像处理在指纹识别中的应用 摘 要 指纹具有唯一性和稳定性，因此被人们用来当作鉴别个人身份的主要依据。随着光学技术、化工技术、纳米技术等多种学科的 快速发展．指纹显现和提取技术取得了较快的控展。但有不少显现或提取得到的指纹效果较差，不易分辨指纹纹线与客体背景主间的差异或指纹纹线成像模糊等，主要表现为指纹纹统与客体背景的反差较弱。指纹纹线受客体背景的干扰、两枚或多枚指纹相互重叠干扰、弯曲表面客体上的指数威像问题等。但由于存在指纹图像的噪声和皮肤弹性等因素影响，指纹识别一直存在识别率不高、运算速度较慢的问题。这时可利用数字图像处理技术对不易辨识的指纹进行增强处理．便于后续的指纹识别鉴定。本文总结了基于小波变换的数字图像处理在指纹图像增强、指纹图像二值化、指纹图像压缩编码、指纹图像细化、指纹图像特征提取等方向的各种算法及技术。另外本文还给出了基于matlab软件的指纹自动识别系统实现。在指纹图像的预处理中，首先进行分块归一化，为后续处理提供统一的规格图像；在求方向图中，用沿着某个方向的灰度方差代替Metre方法中的灰度变化，相当于在求点方向图之前先进行了一次均滤波操作，这样得到的方向图更有鲁棒性；在二值化中，阀值的选取引入最大熵的概念，使图像具有抗噪性。但对于部分噪声严重的指纹图像仍然无法识别，另外，算法的运行效率还有待提高。在指纹图像的降噪中：应用中值滤波与小波包变换相结合去除图像随机噪。 关键词：数字图像，指纹处理，小波变换，matlab，指纹识别系统研究 注：本设计（论文）题目来源于教师的国家级（或部级、省级、厅级、市级、校级、企业）科研项目，项目编号为： 。 Abstract Fingerprint is unique and stability, and therefore are used as main basis of personal identity. With the rapid development of optical technology, chemical technology, nanotechnology and other disciplines. Fingerprint and extraction technology has made rapid development. But many poor fingerprint effect appeared or extract, is not easy to distinguish the difference between background and object of the main ridge or fingerprint image blur, mainly for the contrast fingerprint system and object background of the weak. Interference, fingerprint by object background two or more fingerprints overlap interference, index Wei curved surface objects like problem etc.. But because of the existence of the fingerprint image noise and the elasticity of the skin and other factors, the fingerprint recognition has been the recognition rate is not high, the low speed problem. Then the difficult identification of fingerprint enhancement processing by using digital image processing technique for fingerprint identification later. This paper summarizes the wavelet transform of digital image processing in the fingerprint images enhancement, two values, fingerprint image compression coding, the fingerprint image thinning, fingerprint image feature ext