基于机器视觉的agv动态路径识别算法-研究

《基于机器视觉的agv动态路径识别算法-研究》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、兰州交通大学硕士学位论文AbstractWiththedevelopmentofmodemindustrialautomation,versatileautomatictridimensionalstorehousehasdominatedvariousindustrialfields．Asbeingthelinkandtheterminalthatadjuststheflexiblelogisticsystem_,AutomaticGuideVehicle(AGV)representsaniconicsymbolofmodemcontinuou

2、slogistictransportation．Inordertoenhancetheaccuracy，timeliness，andadaptationofvisualguidingAGVs，toachievemo∞agileguidingandefficientpathplanasthefinalresult,thethesisfocusontheconzi．nm锣ofadjacentframes，analyzedandstudiedthenavigationlines，marksandobstaclesintheproject．Bysmdym

3、gthecurrentdevelopmentof娲＼sinothercountries，ar谢yzingandcomparingtheoperationalprincipleofvariousguiding／navigationsystemamong石G鸭thethesisstatedsuperioritiesandflawsofthevisualguidingAGVsystem,hencegaveaclearviewofwhatneededtobesolvedfrom也ethesis．沥theestablishmentofmathematica

4、lmo&lingonvideocamerasystem,whichincludedthesettingofcoordinatedsystemandanalysisoflanemodule；duringtheimagepretreatment,andbinaryproeessingofgrayscaleimagebasedsequentialimages：andexaminedthemarginofnavigationmarklines，acquiredthedynamicdataofthepathenvironmentduringAGV’Sope

5、ration．InordertoincreasetheaccuracywhileAGVWastrackingtheroute，studiedthetestingmethodofcentrallineofguidemark,andfoundthatconventionalHoughtransformalgorithmtookmuchlongertimewhileextractingtheguidecentralline．Consequently,studiedtheHoughtransfornlalgorithmthatbasedonpretrea

6、tment．，andextractedthecentralguidingline．Astheresultshowed,thatimprovedHoughtransformalgorithmenhancedthetimelinessofentiresystem,provedvalidityofimprovedalgorithm．OnaccountoflackinginenoughdatafromthesimplexnavigationlinesthatinpracticaloperationsofavisualguidingAGV,andforth

7、epurposeofimprovement,enrichthepathinformationandenhancetheintelligenceofthesystem,thereweretwosortsofsimplifiedguidingmarksthatdesignedforthesystemidentifying，theywere：controlmarksanddi百talmarks．Thesetwosortsofmarkswerebothtreatedbyimageprojectionthatprojectedtotheinterested

8、area,thenidentifiedthemarksaccordingtotheirOWllstructuraleigenveetor