资源描述：
《智能小车避障与路径优化研究》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、万方数据AbstractAbstractIntelligentvehiclehasagreatapplicationforegroundinthefieldofthemilitary,spaceexplorationandlogisticsbasedontherapiddevelopmentofcomputertechnology,sensortechnology,aswellasvisiontechnology.Automaticnavigationachievedbymeansofmachinevision
2、hasbeenahotresearchareawhichplaysakeyrole.Theenvironmentinformationgatheredaroundthroughmonocularcameracanbeusedtoextractvisualinformationwhichacquiredbyimageprocessingalgorithm.Simultaneously,obstacleswhichdistributerandomcanbefoundout,whichobviouslycontrib
3、utetohelpdesignafeasibleoptimalpath.Inthispaper,thevisionsubsystemconsistsofmonocularcamera,whichwillbeappliedtostudyautomaticnavigationbasedonmachinevision,whileobstacleavoidanceandpathoptimizationproblemshouldbetheprimarysolution.Thus,therelevantcontentsin
4、cludecameracalibration,obstacledetection,especiallyobstacleavoidanceandpathoptimization.Thesystemstructurebelongtoexperimentalvehiclewillbefirstlydiscussedbriefly.
Referencecoordinatesystemhasbeenintroducedtofocusondescribingimagingmodel
formedfrommonocularc
5、amera,eventuallycameracalibrationwillbefinishedbyprogram
madeupofOpenCV.Consideringdecomposingvideosignalcapturedbycamerainto
imagesequence,whichsuitsfurtherstudy.Thisthesisdetailedlyanalysestheobstacledetectionalgorithmbasedonimage
sequence.Manyshortcomings
6、canbeeasilygainedonceapplyingframedifferentiation
method.ItisprovedthatAverageBackgroundModelandGaussianMixtureModelhave
noadvantagesaswell.Thus,Amulti-featurefusionbasedapproachwillbeputforwardto
overcomeit.Texturefeatureandedgeinformationweremergedwithline
7、arweighted
method,comparingoveralldifferencesbetweentwoadjacentframesonceupdatingthe
backgroundinformationestablishedbytheGaussianmodel,obstaclecanbedistinguished
asexpected.Falseobstacleshaveagreatinfluencetothedetectionresults,Anew
segmentationapproachbase
8、donmaximumentropycanworkhere.What’smorethat
Houghtransformsandmorphologicalprocessingwillhelpidentifyrealobstacles,once
connecteddomaincalibrationhasbeendone,obstaclescanbefoundout.Itsuggeststha
