摘要
Inthispaper,wepresentarobotvisionbasedsystemforcoordinatemeasurementoffeaturepointsonlargescaleautomobileparts.Oursystemconsistsofanindustrial6-DOFrobotmountedwithaCCDcameraandaPC.Thesystemcontrolstherobotintotheareaoffeaturepoints.Theimagesofmeasuringfeaturepointsareacquiredbythecameramountedontherobot.3Dpositionsofthefeaturepointsareobtainedfromamodelbasedposeestimationthatappliestotheimages.Themeasuredpositionsofallfeaturepointsarethentransformedtothereferencecoordinateoffeaturepointswhosepositionsareobtainedfromthecoordinatemeasuringmachine(CMM).Finally,thepoint-to-pointdistancesbetweenthemeasuredfeaturepointsandthereferencefeaturepointsarecalculatedandreported.Theresultsshowthattherootmeansquareerror(RMSE)ofmeasurevaluesobtainedbyoursystemislessthan0.5mm.Oursystemisadequateforautomobileassemblyandcanperformfasterthanconventionalmethods.
出版日期
2016年01月11日(中国期刊网平台首次上网日期,不代表论文的发表时间)