基于LabVIEW的机器视觉实现

:2002-03-05:(1976-),,,,基于LabVIEW的机器视觉实现裴忠发,赵敬斌,罗志增(,310037):介绍了从图像获取图像处理图像理解等几个方面去实现具有视觉功能装配机器人的方法:机器视觉;图像处理;图像理解:TP2426+2:A:1001-4551(2002)04-0053-03RealizingofRobotVisionBasedOnLabVIEWPEIZhongfa,ZHAOJingbin,LUOZhizeng(RobotInstitute,HangzhouInstituteofElectronicsEngineering,Hangzhou310037)Abstract:Thearticledescriblesthemethodofrealizingassemblyrobotsvisionfromtheaspectofimageacquisitionimageprocessandimagecomprehension.Keywords:robotsvision;imageprocess;imagecomprehension1,,2050,(Marr),,,,,NILabVIEW,2LabVIEWLabVIEW,VI(VirtualInstruments,LabVIEW)1VI3:LabVIEW,331(1)CCD,(2)A/D,PC,CCD(3)PCCCDPC,LabView,1图1系统硬件组成结构图32VFW(VideoforWindows)Windows,AVI,(API),VC++532002194Mechanical&ElectricalEngineeringMagazineVol.19No.42002(1)capCreateCaptureWindow,,:hwndC=capCreateCaptureWindow(!myimage∀,WS-TILEDWINDOW|WS-VISIBLE,X1,Y1,X2,Y2,NULL,nID);,nID(2)CapDriverConnect,:capDriverConnect(hwndC,nIndex),hwndC,nIndex(3)FrameCallbackProc,:FrameCallbackProc(HWNDhwnd,LPVIDEOHDRlpVHdr)(4)capFileSaveDIB,:capFileSaveDIB(hwndC,!C:\xxxx.bmp∀),4,,,(1),,S,S:g(x,y)=1M#(x,y)∃Sf(x,y)(1)M%%%S%%%(x,y)(2),,,:G[f(x,y)]=fx2+fy21/2(2),:G[f(x,y)]=f(i,j)-f(i+1,j)2+f(i,j)-f(i,j+1)22(3)(3),,M,rk(k=0,1,2,3,&,M-1),Pr(k):Pr(k)=nk/n(k=0,1,2,&,M-1)(4)nk%%%n%%%,,,(4),,,:!2f=2fx2+2fy2(5),:2f(i,j)x2=∀xf(i+1,j)-∀xf(i,j)=f(i+1,j)+f(i-1,j)-2f(i,j)(6)2f(i,j)y2=∀yf(i+1,j)-∀yf(i,j)=f(i,j+1)+f(i,j-1)-2f(i,j)(7):∀2f(i,j)=f(i+1,j)+f(i-1,j)+f(i,j+1)+f(i,j-1)-4f(i,j)(8)5,:,,A:A=k∋n(9)k%%%n%%%S:S=l2/A(10)54Mechanical&ElectricalEngineeringMagazineVol.19No.420022002194BP神经网络补偿三维曲面非接触式测量系统热变形误差研究傅龙珠,陈亚良,狄瑞坤(,310027):针对一个三维非接触式测量系统,在简要介绍系统工作原理的基础上,指出机械部分的热变形误差是影响系统误差的关键该文在分析测量系统机械部分的基础上对它进行了热误差的分析和建模,并结合改进的BP神经网络给出了具体实现的方法:热变形误差;误差补偿;BP神经网络:TP183;TH161+4:A:1001-4551(2002)04-0055-04StudyonBPNeuralNetworkforThermalDeformationErrorAnalysisandCompenstationofaNoncontact3DMeasuringSystemFULongzhu,CHENYaliang,DIRuikun(InstituteofProductionEngineering,ZhejiangUniversity,HangZhou310027)Abstract:Aimsattheworkprincipleofathreedimensionalnoncontactmeasuringsystem,thispaperpointsoutthattheerrorofmechanicalpartexits.Onthebasisofdesign,thegeometricerrorandthermalinducederrorofthemechanicalpartisanalyzedandmodeled.ThispapergivestheactualmethodaccordingtotheimprovedBPneuralnetwork.Keywords:thermaldeformationerror;errorcompensation;BPneuralnetwork1,,2315mm∋1970mm∋1910mm:2001-12-19:2002-04-23:(1977-),,,,CNCl%%%(p+q):mpq=#x#yxpyqf(x-#x)p(y-#y)qf(x,y)(11)x,y%%%,,,,,,6LabView,LabViewVI,:VIVIVIVIVIVI,VI,,7NILabVIEW,[1]夏良正.数字图像处理[M].南京:东南大学出版社,1999.[2]何发昌,邵远.多功能机器人的原理及应用[M].北京:高等教育出版社,1996.[3]贾云得.机器视觉[M].北京:科学出版社,2000.552002194Mechanical&ElectricalEngineeringMagazineVol.19No.42002