文档介绍：兰恤t学硕if-拳住论文摘要指纹因其唯一性和稳定性被广泛用于身份识别领域。自动指纹识别系统(AFIS)基于图形图像处理、数据库等计算机技术来进行指纹识别,尽管在此技术上已经有多种成型产品,但是许多核心技术因商业利益而未公开,社会的发展对系统的性能也提出了更高的要求,指纹识别领域仍是当前及很长一段时间的研究热点,具有重要的理论意义和使用价值。本文在借鉴前人成果的基础上,完成了以下工作: (I)实现了一个完整的指纹图像预处理及特征点提取过程,具体为分离指纹图像背景、求取指纹方向图、指纹图像二值化(分割)、细化、求取特征点、去除伪特征点等算法。(2)在上述指纹二值化(分割)N指纹分割除噪算法,N神经网络的脉冲传播特性,结合指纹图像的方向信息,N的传播特性仅沿指纹脊线方向传播,取得更好的指纹分割效果。(3)基于遗传算法的指纹特征点匹配,采用了一种新的方法先确定两幅指纹的方向旋转角e,然后仅对指纹偏移d](,dy编码,设计匹配适应度函数,采用遗传算法确定两幅指纹的匹配度。在搜索过程中,相对于传统的将方向旋转角0、 dx、dy混合编码的搜索空间大大减小,从而提高了用遗传算法进行指纹匹配的运算速度。(4)考虑到自动指纹识别系统的广阔应用前景,开发了一个简易但具完整功能的指纹考勤系统,可供中小型单位进行日常考勤记录使用。关键字:指纹识别,特征提取,特征匹配,PCNN神经网络,遗传算法,AFIS 兰掘走学硕士学位论吏 Abstract Fingerprint iswidely used as an important means ofidentityrecognition based on itsexcIusive and stablecharacteristic Automatic Fingerprint Identification System (AFIS)is away ofidentity identificationbased on digitalimage processing,database technology and SO on Although there are many mature AFIS product butmost of which are not public because mercial reasons The research offingerprint identificationisstilI apopularsubject. Thispaper isbased on theachievements 0fourpredecessors,and get some achievements asfollowed: (I)Have realized acomplete processing offingerprint minutia extracting, includingcalculating orientation,segmentation,and reducenoising,thinning, searching minutia anderasing faultminutia. (2)Put forward N based on orientation algorithm forfingerprint work hasproperty ofpulse we use the Orientation offingerprint toenablepulse transmitalong theridgeoffingerprint, which can make good resultoffingerprint segmentation. (3)in fingerprint minutia matching process,ic algorithm was articlefirstproposed one kindofprobability algorithm todetermine two fingerprints thedirectionangles ofrotation 0,SOonlydisplace dxanddy tothefingerprint code, icalgorithms was used todetermine twofingerprints the searchprocess,opposite inthetraditionalgorithms,which coded dx,dyand 0mixed, thisalgorithm reduced searchspacegreatly