基于颜色和纹理的图像检索

基于颜色和纹理特征的图像检索摘要随着多媒体信息的发展，基于内容的图像检索（CBIR）逐渐成为一个非常活跃的研究领域，但是目前较为常见的CBIR系统仍是基于单一特征的图像检索。本文在对颜色特征和纹理特征的研究基础上，将两种特征结合在一起，实现了一种综合特征的检索系统。图像特征的提取和相似性度量是CBIR的两个关键技术。由于灰度共生矩阵能够反映图像纹理中灰度级的空间关系，并且计算速度快，因此本文对纹理特征的提取采用灰度共生矩阵的方法。由于HSV空间更符合人类视觉系统，本文采用在HSV空间中提取颜色直方图的方法表示颜色特征。为了进一步提高检索准确率，对颜色特征提取之前进行分块处理。通过实验说明本文中所提出的这种综合检索方法能够有效的利用颜色和纹理特征，检索出两种特征均与检索样图相似的图像。关键词：图像检索共生矩阵HSV空间相似性度量TheImageRetrievalBasedonColorandTexturefeatureAbstractWiththedevelopmentofmultimediainformation,content-basedimageretrieval(CBIR)hasbecomeaveryactivefieldofstudy,thoughmorecommonCBIRsystemisstilltheimageretrievalbasedonsinglefeature.Onthebasisofstudyonthecolorfeaturesandthetexturecharacteristics,inordertoachieveacomprehensiveretrievalsystem,thetwofeaturesarecombinedtogetherinthepaper.TherearetwokeytechniquesinCBIR,whichareimagefeatureextractionandsimilaritymeasure.Becausethespacerelationingrayscaleofimage’stexturecanbereflectedingreylevelco-occurrencematrix,andthecomputingspeedtogreylevelco-occurrenceisfast,thetexturefeatureisextractedbythismethod.BecausetheHSVspaceismorefittothehumanvisualsystemthanothers,thecolorcharacteristicisexpressedinthemethodofextractingcolorhistograminHSVspace.Inordertoimproveretrievalaccuracyfurther,themethodofdoingregionprocessingforimagesbeforeextractingfeaturesisusedinthispaper.Theresultsofexperimentsshowthatcolorfeaturesandtexturefeaturescanbeusedefficientlyandthoseimageswhichhavesimilarcolorandtexturefeaturestothesampleimagecanberetrievedinthismethod.Keywords:imageretrieval;colorcharacteristics;texturefeature;similaritymeasure目录1引言................................................................................................................................11.1CBIR研究的背景和意义......................................................................................11.2CBIR的研究现状.................................................................................................11.2.1基于颜色特征的图像检索.........................................................................21.2.2基于纹理特征的图像检索.........................................................................31.2.3基于形状特征的图像检索.........................................................................31.3CBIR的特点.........................................................................................................41.4CBIR检索方法的分类..........................................................................................41.4.1外部图例查询.............................................................................................41.4.2内部图例查询.............................................................................................41.4.3草图查询.....................................................................................................51.4.4综合检索方法.............................................................................................51.5CBIR系统的分类.................................................................................................61.5.1基于特定图像数据库的图像检索系统.....................................................61.5.2基于网络的图像检索系统.........................................................................61.6CBIR的研究方向.................................................................................................61.7CBIR应用领域......................................................................................................72系统设计........................................................................................................................82.1系统的基本原理及框架.......................................................................................82.2系统的关键技术...................................................................................................92.2.1图像特征的提取.........................................................................................92.2.2基于特征的相似度匹配...........................................................................102.3图像检索的性能评价.........................................................................................123基于纹理特征的图像检索..........................................................................................133.1纹理的简介.........................................................................................................133.2纹理特征的提取方法.........................................................................................133.2.1灰度直方图的矩.......................................................................................133.2.2灰度共生矩阵...........................................................................................133.3基于灰度共生矩阵的图像检索.........................................................................173.3.1基于灰度共生矩阵的图像检索流程.......................................................173.3.2实验结果及分析.......................................................................................184基于颜色特征的图像检索..........................................................................................214.1颜色的简介.........................................................................................................214.2颜色空间......................................