基于matlab的指纹图像特征提取.doc

单位代码学号分类号密级毕业设计(论文)基于Matlab的指纹图像特征提取院（系）名称专业名称学生姓名指导教师2012年5月15日第PAGEIV页基于Matlab的指纹图像特征提取摘要随着社会的发展，传统的基于信物或口令的安全系统显得越来越脆弱，不能适应现代安全系统的需要，因而人们需要研究更加安全可靠，防伪性能更好的安全系统。指纹识别技术就是在这种背景下产生的，它借助人体的生理特征来提高身份识别的可靠性，目前已经成为国内外研究的热点。指纹识别大体分为三个步骤：预处理、特征提取和特征匹配。本文集中于研究特征提取部分，并针对特征提取中的一些关键算法和实现进行了研究和优化，其主要内容如下：在特征提取方面，本论文采用了一种8邻域编码纹线跟踪算法，标注出端点和分叉点来进行特征提取；在剔除伪特征点时，先进行去边缘处理，再根据不同类型伪特征点的特征，采用相应剔除算法。实验表明，以上算法具有较小的运算量和较高的准确性。上述算法在本文中均用Matlab实现，取得了较好的效果，为后续的特征匹配工作打下了良好的基础。关键字：指纹特征提取，去除伪特征，算法仿真Matlab-basedfingerprintimagefeatureextractionAbstractWiththedevelopmentofsociety,thetraditionalsafetysystembasedonkeepsakeandpasswordhasbeenweaker,Whichcannotmeettherequirementofmodernsafetysystem.Inthiscase,theneedofamorereliablesafetysystemwithhigheranti-fakeperformancepromptstheappearanceoffingerprintidentificationtechnique.Thistechnique,withahighersafetyandreliableperformance,canimprovethereliabilityofidentityresolutioninvirtueofhumanbody’sphysiologicalfeature,andithasbeenaresearchfocusthesedays.Fingerprintidentificationfallsintothreeparts,theyarepretreatment,featureextraction,andcharacteristicmatching.Thethesismainlyfocusfeatureextraction,itoptimizedandinnovatedsomekeyalgorithmsofthisparts,whichcanbedescribedasfollows:inthefeatureextractionpart,thethesisusedaeight-neighborhoodcodingridgetracingalgorithm,removingsometemplatesofconsecutivepointsandbifurcatepointswhichhavebeenoptimizedandremovedinthethinningalgorithm,andfinallymarkingterminatepointsaswellasbifurcatePointstoexecutefeatureextraction.Experimentresultindicatedthatsuchnewalgorithmhasalessoperationbutwithahigheraccuracy.AllthealgorithmsintroducedabovehavebeenimplementedonMatlab,andresultprovedanadaptivegoodeffect,whichfacilitatesthenextcharacteristicmatchingprocess.Keywords:Featureextraction,Removingoffalsecharacteristicpoints,Algorithmsimulate目录TOC\o"1-3"\h\z\uHYPERLINK\l"_Toc323801528"1绪论PAGEREF_Toc323801528\h1HYPERLINK\l"_Toc323801529"1.1引言PAGEREF_Toc323801529\h1HYPERLINK\l"_Toc323801530"1.2生物识别技术简介PAGEREF_Toc323801