基于双目视觉移动机器人的路径规划和避障研究

54622009,30(23)ComputerEngineeringandDesign0[1-3][4-5][6-8]112008-12-292009-02-15(J50103)(1984)(1961)[e-mail](200072);;;;;TP391.4A1000-7024(2009)23-5462-05ResearchofpathplanningandobstacleavoidancemobilerobotbasedonbinocularvisionZHAOJing,CHENYi-min(SchoolofComputerEngineeringandScience,ShanghaiUniversity,Shanghai200072,China)AbstractAsystemdesignschemeofmobilerobotisproposed.Themobilerobotimplementsthefunctionofpath-planningandobstacleavoidance.Thepaperexplainshowtopercepttheenvironmentbybinocularvision,howtoseparatetheground,backgroundandobjectsbyusingacombinationofseveralimageprocessingalgorithm,andhowtodistinguishtheobstaclesandtargetbyimagestatisticalchara-cteristics.Atthesametimeweplanthepathbyimprovingartificialpotentialfield,anddesignafuzzycontrollerandformulatingase-quenceofobstacleavoidancerulesbasedonfuzzylogic.Theresultofthesystemprovesthatthedesignisfeasibilityandvalidity.Themobilerobotsystemrealizetherequestoftheproject:inthehelpofbinocularvisionsensors,planningapathdynamicallyandavoidingtheobstacles.Keywordsmobilerobot;binocularvision;imageprocessing;momentinvariants;artificialpotentialfield;fuzzycontrolComputerEngineeringandDesign2009,30(23)54632MicrosoftDirectShowDirectShowIntelOpenCV(opensourcecomputervisionlibrary)OpenCVDirectShowpSampleOpenCVIplImageOpenCVIplImage[9-10]PPdP=/+(1)+ff(1)233.13.2[11]NCCNCC312(x,y)yd(25,350)d25ord350NNYd350F25Y54642009,30(23)ComputerEngineeringandDesign,==1=1,,,×,=1=1,,2×=1=1,2(2)(2)(2),,,,=,=1=1,,,=1=100101(2)T32+12,=,+1,+,11,12,=2,+21,+2,121,1,0,,=2+1,+121,+12+1,1+21,1=1=1,,,233.3[11](p+q)=++,(,=0,1,2,,)(M×N),==1=1,p+qn7XY,=||||==12n=7XY,,44.34.1[12]Khatib=2,k,,=||||3(b)(a)+111+1NN2009,30(23)5465==,=1,10,,00,,0,0==1,1012,,,00,,0=+*=121,1022,,00,,0(3),,==1+2,,00,,0(4)1=(1,10),2,(5)2=12(1,10)21,(6)1260112=1(,)0(,)0(5)(6)1=(1,10),2,2=12(1,10)21212(1,10)21(3)4.233[,,]4,10~100cm{(),()}-90°~90°{LB(),LS(),Z(),RS(),RB()},4(a)4(b)515385404.3(a)4,1030507090NearFar10.80.40.20(b)0°30°60°80°120°LBLSZRSRB10.80.40.20-120°-90°-60°-30°1R1R2R3R4R5farfarfarfarfarnearnearnearnearnearfarfarfarfarfarNBNSZRSRBNBNSNSRSRB54662009,30(23)ComputerEngineeringandDesign(1)(2)7(3)-90°~90°sign_len-gthsign_length=sign_length/20Lsign_lengthLsign_length55(a)5(b)6(a)6(b)6(b)7(a)7(b)7(a2)7(a1)7(a2)7(b)8(a)8(d)6NCC:[1],,,.[J].,2008,26(2):119-123.[2],.7-DOF[J].,2005,17(8):1948-1950.[3].[J].,2002(1):10-12.(5470)5(a)(b)7(a)(b)(1)(2)21348(a)(b)(c)(d)6(a)(b)54702009,30(23)ComputerEngineeringandDesign5LSBLSB23[5]:[1]KerAD.SteganalysisofLSBmatchingingrayscaleimages[J].IEEESignalProcessingLetters,2005,12(6):441-444.[2]HolotyakT,FridrichJ,VoloshynovskiyS.Blindstatisticalstega-nalysisofadditivesteganographyusingwavelethigherorderstatistics[C].LNCS3677,2005:273-284.[3]GoljanM,FridrichJ,HolotyakT.Newblindsteganalysisanditsimplications[C].ProceedingsofSPIE,Security,Steganography,andWatermarkingofMultimediaContents,2006:1-13.[4]MarvelL,HenzB,BonceletC.Aperformancestudyof±1stega-nalysisemployingarealisticoperatingscenario[C].ProcofMIL-COM,2007.[5]JunZhang,CoxIngemarJ,DoerrGwenael.SteganalysisforLSBmatchinginimageswithhigh-frequencynoise[C].ProceedingoftheIEEEMMSP,2007:385-388.[6]TheUSDANRCSphotogallery[DB/OL].[7]HarmsenJ,PearlmanWA.Steganalysisofadditivenoisemode-lableinformationhiding[C].SanJose,CA:ProcoftheSPIE,Se-curity,Steganography,andWatermarkingofMultimediaCon-tentsVI,2003:131-142.[8],.MATLAB[J].,2006,23(1):131-132.53ROC(10%100%)/%(AUC=0.5609)Zhang(AUC=0.5414)0102030405060708090100Ker(AUC=0.5317)1009080706050403020100/%(a)10%/%(AUC=0.9673)Zhang(AUC=0.9404)02468101214161820Ker(AUC=0.9059)1009080706050403020100/%(b)100%13TP50%80%FP/%(10%)TP/%5080KerZhang4544.440.178.377.57523TP50%80%FP/%(100%)TP/%5080KerZhang6.41.41.414.43.82.5[4]Cog[EB/OL].[5]CASIA-I[EB/OL].[6][EB/OL].=46.[7],.[J].,2007,27(6):378-379.[8][EB/OL].[9]ZhangHandong,WangGang,CenYuwan.Summaryofdistancemeasurementbasedonvisioninlocalizationtechnology[J].PowerElectronicsandMotion,2006,3:1-5.[10],,.[J].,2005,23(1):55-59.[11],,.[J].,2005,17(10):179-187.[12],.[J].,2007,14(2):2152-2156.(5466)