图像分割技术与MATLAB仿真

中南民族大学毕业论文(设计)学院:计算机科学学院专业:自动化年级:2012题目:图像分割技术与MATLAB仿真学生姓名:高宇成学号:2012213353指导教师姓名:王黎职称:讲师2012年5月10日中南民族大学本科毕业论文（设计）原创性声明本人郑重声明：所呈交的论文是本人在导师的指导下独立进行研究所取得的研究成果。除了文中特别加以标注引用的内容外，本论文不包含任何其他个人或集体已经发表或撰写的成果作品。本人完全意识到本声明的法律后果由本人承担。作者签名：年月日目录摘要..................................................................................................................................................1Abstract.............................................................................................................................................1引言..................................................................................................................................................31图像分割技术...............................................................................................................................31.1图像工程与图像分割.......................................................................................................31.2图像分割的方法分类.......................................................................................................42图像分割技术算法综述...............................................................................................................52.1基于阈值的图像分割技术...............................................................................................52.2边缘检测法........................................................................................................................52.3区域分割法.......................................................................................................................72.4基于水平集的分割方法...................................................................................................82.5分割算法对比表格...........................................................................................................83基于水平集的图像分割................................................................................................................93.1水平集方法简介...............................................................................................................93.2水平集方法在图像分割上的应用...................................................................................93.3仿真算法介绍.................................................................................................................103.4实验仿真及其结果.........................................................................................................11结论................................................................................................................................................18致谢..............................................................................................................................................19参考文献.........................................................................................................................................191图像分割技术研究及MATLAB仿真摘要：作为一项热门的计算机科学技术，图像分割技术已经在我们生活中越来越普及。顾名思义这项技术的目的就是，将目标图像从背景图像中分离出去。由于这些被分割的图像区域在某些属性上很相近，因此图像分割与模式识别以及图像压缩编码有着密不可分的关系。完成图像分割所采用的方法各式各样，所应用的原理也不同。但他们的最终目的都是把图像中性质相似的某些区域归为一类，把性质差异明显的不同区域分割开来。通常在分割完成之后，我们就要对某些特定区域进行分析、计算、评估等操作，因而分割质量的好坏直接影响到了下一步的图像处理[1]，因此图像分割是图像处理的一个关键步奏。图像分割技术在各个领域都有着及其重要的意义；在工业上有卫星遥感，工业过程控制监测等等；在医学方面，水平集的分割方法还可以通过医学成像帮助医生识别模糊的病变区域；在模式识别领域还可应用到指纹扫描、手写识别、车牌号识别等等。本课题的研究内容是对图像分割技术的几种常用的方法进行综述和比较，并基于其中一种方法进行MATLAB仿真测试，给出性能分析比较结果。关键字：图像分割，MATLAB仿真，模式识别ImageSegmentationandMatlabSimulationAbstract:Imagesegmentationistoimagerepresentationforthephysicallymeaningfulregionalconnectivityset,namelyaccordingtothepriorknowledgeoftargetandbackground,weontheimageoftargetandbackgroundoflabelingandlocalization,thenseparatetheobjectfromthebackground.Becausethesesegmentedimageregionsareverysimilarinsomeproperties,imagesegmentationisoftenusedforpatternrecognitionandimageunderstandingandimagecompressionandcodingoftwomajorcategories.Becausethegeneratedinthesegmentedregionisakindofimagecontentrepresentation,itistheimageofvisualanalysisandpatternrecognitionbasedandsegmentationresultsofqualityofimageanalysis,recognitionandinterpretationofqualityhasadirectimpact.Imagesegmentationitisaccordingtocertainfeaturesoftheimage(suchasgraylevel,spectrum,texture,etc.)toacompletepictureoftheimageissegmentedintoseveralmeaningfularea.Thesefeaturesmadeinacertainregionofconsistentorsimilar,andbetweendifferentregionsshowedsignificantlydifferent.Imagesegmentationtechnologyinvariousfieldshavemostofthefieldanditsimportantsignificanceindigitalimageprocessing,imagesegmentationhasawiderangeofapplications,suchasindustrialautomation,processcontrol,onlineproductinspection,imagecoding,documentimageprocessing,remotesensingandmedicalimageanalysis,securitysurveillance,aswellasmilitary,sportsandotheraspects.Inmedicalimageprocessingandanalysis,imagesegmentationforbodyoccurrenceofthree-dimensionaldisplayofthediseasedorganorlesionlocationdeterminationandanalysisplaysaneffectiveroleincounseling;intheanalysisandapplicationofroadtrafficconditions,2availableimagesegmentationtechniquefrommonitoringoraerialfuzzycomplexbackgroundseparationtoextractthetargetvehicle.Theresearchcontentofthispaperist