Study On Vision System Of Gluing Manipulator Based On Labview

The manufacture of sanitaryware is being automatic with the development of the science and technology, such as the high-pressure molding, glazing, drying and firing of closestool. The high-pressure molding of closestool produces two parts, the high part and the low part. The gluing which puts the high part and the low part together has been accomplished by the hands for a long term, but its environment is bad, quality is poor and producing efficiency is low. So gluing by manipulators instead of handwork can promote the establishing of the automatic product line, which is the trend of the sanitaryware industry.A machine vision system to get the position information based on LABVIEW for the fieldwork of closestool manufacture in Tangshan Huida Sanitaryware Group was built, considering the big area, anomalous shape and no obvious anchor surface in the high part and the low part of the closestool and the requirement of painting and gluing manipulators. This system should get the painting track coordinates, the position coordinates and orientations for the painting and gluing manipulators.It gets the position information from two cameras which respectively correspond with the high part and the low part in two conveyor belts, by analyzing the requirement of the manipulators. Then, transmitting the position information into PLC to make it control the two manipulators. The hardware of the machine vision system was selected, including CCD cameras, industrial fixed focus camera lens, image acquisition card, lamp-house and cables, then the image acquisition software was programmed.Because lots of noises can be brought in by its bad environment, different kinds of edge detection arithmetic were studied, contrasted, analyzed and experimented to obtain the high precision and quick processing speed to the image of the high part and the low part of closestool. Multiscale wavelet edge detecting arithmetic can take advantages of much resolving capability, ensure edges in the big scales and select edges in the little scales, so it has the perfect edge detection capability. The edges were detected by multiscale wavelet, which contains designing wavelet function, ensuring the threshold value and scales, fusing multiscale images.The machine vision software was designed and programmed, which contains image pretreatment, multiscale wavelet edge detection, the selection of object, the acquirement of painting track coordinates, and the ensuring of gravity center and principal axis orientation. In the end, data was transmitted to PLC to control the manipulators.