【作者】 吴黎娜；

【导师】 侯建军；

【作者基本信息】 北京交通大学 ， 电路与系统， 2008， 硕士

【摘要】 指纹识别技术以生物特征为基础,以信息处理技术为手段,将生物技术和信息技术有机结合在一起,在所有包含身份认证的领域均有着广阔的应用前景。尽管指纹识别的理论研究与应用开发取得了重大进展,但指纹识别算法的准确性和识别速度仍有待提高,构建准确可靠、安全实用的自动指纹识别系统已成为研究热点。本文对指纹识别算法和硬件平台实现两方面展开了较为深入的分析与研究,提出了指纹识别算法的优化方法,并在此基础上设计合理的DSP(Digital SingleProcessing)硬件解决方案。本文的主要目的是将指纹识别技术应用于身份认证领域,促进指纹识别技术在实际应用中的发展。论文的主要研究工作如下:1.研究分析了国内外自动指纹识别技术应用及发展现状。重点研究了指纹识别算法的指纹图像预处理、指纹特征提取和指纹图像匹配三个阶段,比较分析了现有算法的优缺点。2.分析及优化了指纹识别算法中的重要环节。在指纹图像特征提取阶段,为进一步简化图像,提出将经典细化后的图像与8连通模板比较的方法,实现了图像的进一步简化,减少了跟踪过程的难度。在细节点匹配阶段,通过对指纹图像进行分类,对基于三角形全等的指纹细节点匹配的算法进行了优化,减少了匹配过程的计算量。3.设计了一个基于DSP的便携式指纹识别系统硬件平台。指纹识别系统硬件平台设计包括:以DSP芯片TMS320VC5416和指纹传感芯片MBF200为核心,外配片外SRAM和FLASH存储器以及LCD显示屏等的自动指纹识别系统,对各模块进行合理的组合与设计。通过对指纹识别系统的研究,提出了指纹图像预处理及匹配算法的优化,仿真实验结果表明,本文提出的优化算法是可行的。设计了一种具有通用性的基于DSP应用的指纹识别系统,作为该指纹识别算法的应用平台,可有效促进指纹识别技术的应用和普及。更多还原

【Abstract】 Fingerprint identification technology based on biological characteristics, with information processing technology as a means, combines biotechnology and information. It has wide application prospect in all fields which contain identity authentication. Although theoretical research and application development of it have got a significant progress, accuracy of the algorithm and processing speed still need to be improved. Construction of accurate and reliable, safe and practical automatic fingerprint identification system (AFIS) has become a research hotspot.In this paper, fingerprint identification and hardware design are analyzed and discussed in detail, putting forword the optimized algorithm and hardware design based on DSP. The main purpose of this study is to use the fingerprint identification technology for identity authentication, optimize fingerprint identification algorithm, and put forward a DSP scheme. This paper is made up of the following parts:1. The domestic and overseas actuality of fingerprint identification technology are researched and analyzed. Fingerprint image pretreatment, Minutiae Extraction and fingerprint matching is reserarched with emphasis.2. The main phases of fingerprint identification algorithm are improved and optimized. In the process of minutiae extraction, classical thinning images and 8-connected templates are complared to reduce the difficulty of tracking. In the processing of image matching, the imge is assorted before the triangle-based fingerprint minutiae matching algorithm, reducing the calculation greatly.3. The hardware design of fingerprint identification introduction is using digital signal processing TMS320VC5416 and fingerprint sensor MBF200 as its main chips, with the cooperation of SRAM memory, FLASH memory, LCD and so on.This research project has done an optimized research on fingerprint identification algorithm and provided a DSP-application-based fingerprint process module of general type as an application platform for fingerprint identification algorithm. Simulation has proved that optimized algorithm presented by this paper is practicable. This research projection can promote the application and popularization of this technology.更多还原