基于改進遺傳算法的攝像機標定技術的研究

1、大連理工大學碩士學位論文基于改進遺傳算法的攝像機標定技術的研究姓名：劉俏申請學位級別：碩士專業(yè)：機械電子工程指導教師：邢英杰20071223大連理工大學碩士學位論文ResearchOllCameraCalibrationBashonImprovedGeneticAlgorithmAbstractVisionisthemainapproachforpeopletoapperceivetheconditionwordandtoknowthe

2、outsidewordWiththedevelopmentofcompmertechnologyandtheincreasingrequirementfordigitalinfo，computervisionandbinocularstereovisiondevelopgreatlyResumedfrom2Dinlageto3Dgeometrystructure(named3DReeo刪on)isthemainpurposeofcomp

3、utervisionCameracalibrationisthepivotalstepin3DReconstructionandthepremiseandbasementofobtaining3DspacⅪinformationbycomputervisionalsothemainpartofbinocularstereovisionThcmainworkthispaperistoresearchthetechnologyandthem

4、ethodrelationtothec81n@racalibrationtoprovidereliabledamandlayagoodfoundationfornextstepofcomputervisionandbinocularstereovisionbyobtainingtheparametersofcaineraand=rndnguptherelationbetw∞n3Dspaceand2Dinlage，rKspaperdesc

5、ribestheachievementsinbinocularstereovisionathomeandabroadindetallbasedonMart’scomputervisionrationale11圯focusisonresearchingcameracalibrationmethodandtheeadyfeaturepointdetectionprocessFirstresearchimagefeamrepoiⅡtex缸ll

6、etionissuesbythenumbersIntheprocessofcomerdeteetioIl，theelements，themethodsandthecapabilityofSUSANoperatorMICoperatorandHarrisoperatorareanalyzedandcontrasted，thenusingVC60tocompletedifferentimages’Harriscomerdetectionn帕

7、detectionresultsshowedthatthespeedofHarriscomerdetectionmethodwasfastandthismethodcangetevenandaccuratecomersSothisalgorithmwasstableSecondly，researchcanleracalibrationmethodandtherelatedtheoryTheprincipleandtheresearchs

8、tatusquaathomeandabroadofcalibrationtechnologyinstereovisiona∞describedindetailThentheexistingcalibrationtechnologyandthedifficultiesincalibrationtechnologyatpresentageanalyz缸InthispapertheGeneticAlgorithmisusedingamerac

9、alibrationtechnology，andthenanalyzesthedisadvantageandshortageintraditionalGeneticAlgorithmAgainstthesedisadvantagesandshortages，animprovedmethodisputforward—addingaselfcorrectingsearchingintervalanditconquersthetraditio

10、nalGeneticAlgorithm’sdisadvantageswhenusinginhighdimensionspaceoptimizationIncrossoverprobabilityandmutationprobability，adaptivecrossovel“andmutationprobabilityisadoptedanditconquersthedifficultiesinconvergence11∞improve