Intelligent Valve Assembly Technology Research Robot Vision

In the production life, Ball valve is a necessary part with wide application,heavy demand, occupying an important position in national economy and theassembly is the post procedure of ball valve production, with an importantposition in the production of ball valve. At present, the domestic enterprises arestill using the traditional human eye recognition method by the skilledworkers in ball valve assembly. This approach is influenced by sense ofresponsibility, technology level and degree of fatigue of workers, and it alreadycan not satisfy automatic, intelligent, high precision, and high efficiencydevelopment of modern manufacturing. So it becomes inevitable to search for avision automatic assembly technology under good technology developmentplatform.Since the advent of machine vision technology, it is continuously applied tovarious occasions, and achieving rapid development. The machine visionsystem is the industrial vision system when used in industrial production; it canbe roughly divided into the detection, location and assembly by using differentfunctions. In this paper, it uses machine vision technology in ball valveassembly, developing ball valve intelligent assembly system by robot based onvision. In the system, the image acquisition of parts by CCD camera is acted as“perception" stage of the intelligent assembly; control processing programwritten is acted as "decision" stage of the intelligent assembly; grippergrabbing is acted as "performing" stage of the intelligent assembly.In this paper, ball valve intelligent assembly system by robot based onvision is studied; the whole system is composed by four main units ofhuman-computer interaction, assembly workbench, vision and robot clampingoperation and so on. It designs visual subunit that composed of two CCDcameras and industrial camera, clamping operation subunit that composed ofpneumatic flexible universal gripper etc; at the same time, it writes the imageacquisition, processing, and camera calibration system program, etc. First of all,the visual subunit captures the image of target parts, proceeding image preprocessing including the wavelet denoising, enhancement, binarization andcomputation of mathematical morphology; Then through methods such astemplate matching and edge detection, it distinguishes and positions the targetparts; Finally through LabVIEW processing program, it gets the poseinformation of parts in local coordinate namely calibration system, and thencontrol flexible gripper to complete the work of assembly and grab.Through field experiment, it shows that the developed visual assemblysystem not only meets the assembly accuracy, but also raises the qualificationrate of products and production efficiency, and reduces the requirement andwork intensity of operator.