基于全景视觉的移动机器人视觉系统研究硕士学位论文

《基于全景视觉的移动机器人视觉系统研究硕士学位论文》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、江南大学硕士学位论文基于全景视觉的移动机器人视觉系统研究姓名：刘光宇申请学位级别：硕士专业：检测技术与自动化装置指导教师：刘国栋20090501AbstractRobotsareautomaticallydesignatedworkmachineequipment．Inaccordancewitllthesurroundingenvironmentinformation，itmaycompletesomespecifictasks．皿1iswillpeoplefromdangerorcomplexworkfreed．删lemobile

2、robottoperceivetheexternalenvironment，themaintoolisvisualsensor．Butthevisualangleoftraditionalcamerasystemhaverestriction,oilendoesnotmeettherequirements．TKspaperintroducesakindofomni．directionalvisiondevicethatllsescameraandthemirrorimageofeffectiveenvironmentaroundaro

3、bot，whichCanprovideinformationoftherobot’s360。space．Itfoundthebasesofself-localizationandnavigation．Tmspapermainlystudiesthedesignofhardwareandsoftwaredevelopmentbasedontheomni—directionalvisionsystemofrobot．，nleexperimentalresultsverifiedtheeffecfivenessofthemethod．Har

4、dware，thispaperstudiestheprincipleofomni-directionalvisionsystem，introducestheinfluenceofsingleviewpoint,comparestheeffectsoffourdifferentshapesofthereflector．Anddeterminetheimagingfeaturestomakefulluseofthehypoidmirrormirror．Onthisbasis，wedesignanornni—directionalvisio

5、nsystem，whichmadeupofanisomerichorizontalmirrorandanisomericverticalmirror,whicheriectivelylimitingomni—directionalvisionimagesdiscrepancy．Software，firstintroducedtheom．ni-directionalvisionsystemprocessingflow,andresearchthecameracalibration，imagecollection，ere．Andbased

6、ontheomni—directionalvisionsystemimages，thepaperdiscusstheimagesmoothandimagesharpenoftheimagepreprocessingprocedure．Secondly,thearticlestudiedthebasicmethodofcolorimageprocessing，introducesthecompositioncolorspaces，comparesthediffenencebetweenseveralcolorspaces．Thepape

7、ralsoimprovestheconversionformulasfromRGBtoⅥⅣcolorspace，studiesthebasicmethodofimagesegmentation．Thenthepaperusesthemethodofmulti-thresholdimagesegmentationofⅥⅣcolortosegmentthecolorimage．Afterthat，inordertocalculatetheregionalcharacteristicsoffeaturepointcoordinates，th

8、earticalrecogniseimageregionalcharacteristics，then，removesinterferenceusingbinarymorphologicalmethod．Finally,s