基于内容的图像检索系统

河南理工大学毕业设计（论文）说明书I摘要近年来，随着多媒体技术和计算机网络的飞速发展，全世界的数字图像的容量正以惊人的的速度增长。这些数字图像中包含了大量有用的信息。然而，由于这些图像是无序地分布在世界各地，图像中包含的信息无法被有效的访问和利用。一般的文字搜索引擎对于图像数据来说，几乎束手无策，这就要求有一种能够快速而且准确的查找访问图像的技术，也就是所谓的图像检索技术。基于内容的图像检索(Content-BasedImageRetrieval，CBIR)指的是查询条件本身就是一幅图像，它建立索引的方式是通过提取图像特征，然后通过计算比较这些特征和查询条件之间的距离，来决定两幅图像的相似程度。本文主要对基于内容的图像检索系统进行设计，系统设计的重点在于图像的低层特征提取。本系统设计主要包括：图像低层特征（颜色、纹理等特征）的提取及描述，多特征融合技术和特征匹配技术等方面。本文首先论述了国内外基于内容的图像检索技术的研究现状、检索特点以及应用领域，然后从图像检索系统的基本原理入手，实现了图像的预处理，图像颜色特征、纹理特征的提取，以及图像库、特征库的构成，接着通过欧式距离进行相似度匹配。最后，本系统为查询接口提供一个友好的用户界面，采用MATLABGUI(GraphicalUserInterface)设计实现了一个基于内容的图像检索系统，并对系统的各个模块及其功能进行了介绍。通过仿真实验，对Corel图像库中的500幅图像进行了检索实验。实验结果表明，本系统计算简单、有效，能够取得较好的检索效果。关键词：图像检索；颜色特征；纹理特征；相似性度量；GUI设计河南理工大学毕业设计（论文）说明书IIAbstractInrecentyears,withthedevelopmentofmultimediatechnologyandtherapiddevelopmentofcomputernetwork,theworldwidedigitalimagecapacityisincreasingatanalarmingrate.Thesedigitalimagecontainsalotofusefulinformation.However,theseimagesarerandomlydistributedthroughouttheworld,theimagecontainsinformationcannotbeaccessedandusedeffectively.Generaltextsearchenginefortheimagedataseemsalmostincapable,thisrequiresafastandaccurateimagesearchingtechnology,alsoknownastheimageretrievaltechnology.Contentbasedimageretrieval(Content-BasedImageRetrieval,CBIR)itselfisapairofimages,itisindexedbytheimagefeatureextraction,thendeterminethedegreeofsimilarityoftwoimagesthroughthecalculationandcomparisonofthefeaturesandthedistancebetweenthequeryconditions.ThispaperfocusesontheCBIRsystemdesign,.Thisdesignisfocusedonlow-levelfeatureextraction.ofimage.Thesystemdesignmainlyincludes:lowlevelimagefeatures(color,texturefeatureextractionanddescription)extractionanddescription,multiplefeaturesfusionandfeaturematchingtechnologyandsoon.Thispaperfirstdiscussesthedomesticandforeigncontentbasedimageretrievaltechnologyresearchstatus,characteristicsandapplicationfieldofretrieval,andthenfromtheimageretrievalsystembasicprinciple,realizationofimagepreprocessing,imagecolorfeature,texturefeatureextraction,andimagelibrary,characteristicdatabase,andthenthroughtheEuclideandistancesimilaritymatching.Finally,thesystemforthequeryinterfaceprovidesafriendlyuserinterface,usingMATLABGUI(GraphicalUserInterface)designandimplementationofacontent-basedimageretrievalsystem,andofthevariousmodulesofthesystemanditsfunctionareintroduced.Throughthesimulationtest,theCorelimagedatabaseof500imagesretrievalexperiment.Theexperimentalresultsshowthat,thissystemhastheadvantagesofsimplecalculation,effective,canobtaingoodretrievalresults.Keywords：Imageretrieval;Colorfeature;Texturefeature;Similaritymeasurement;GUIdesign河南理工大学毕业设计（论文）说明书III目录摘要...................................................................................................................................IAbstract.............................................................................................................................II目录............................................................................................................................III1绪论..............................................................................................................................11.1选题背景及研究意义.......................................................................................11.2图像检索技术发展历程...................................................................................11.2.1基于文本的图像检索............................................................................11.2.2基于内容的图像检索系统....................................................................21.3CBIR技术应用..................................................................................................31.4经典CBIR系统介绍........................................................................................41.5论文结构...........................................................................................................62图像特征分析..............................................................................................................82.1图像颜色特征分析...........................................................................................82.1.1颜色空间及转换....................................................................................82.1.2颜色量化..............................................................................................102.1.3颜色特征的提取方法...........................................................................112.2图像纹理特征分析.........................................................................................132.2.1图像纹理特征描述...............................................................................132.2.2纹理特征的提取方法...........................................................................143特征提取算法选择....................................................................................................193.1颜色特征提取算法.........................................................................................193.1.1颜色的量化..........................................................................................193.1.2颜色特征的提取........................................