Study On Some Key Techniques Of The Laser Scanning Seam Tracking Sensor

The weld is important in people′s life.With the improvement of economical level,people come up with higher requirement about the quality of weld.In order to improve the quality of weld,we must reduce the effect of human factor,and accelerate the application of robot on the weld.The robot needs to be installed on laser scanning seam tracking senor to realize widely automatic weld.The main function of sensor is to obtain the precise position of weld.At present,the arc sensor and the visual sensor are mainly used for seam tracking.With the development of image′s processing technology,visual sensor is more widely used.The laser scanning seam tracking sensor adopted by this system is one kind of visual sensor.It is made up of industrial camera,laser generator,a lighter filter and splash baffle.In this article,some image processing methods were studied include image enhancement,threshold segmentation,lone point filtering and digital image morphology.A series of image processing methods were adopted on the process of laser image to get the pixel coordinate of weld feature points.We study the way of the whole system calibration include camera,linear structure light and robot′s hand-eye,and come up with new laser′s and robot′s hand-eye′s calibration way.Finally we obtain transformational relation of feature point between every coordinate system.Finally,we calibrated robot laser welding seam tracking system and adopted calibrated system to detect and track known weld.Experiment tests that the ways of the process of laser image and the calibration of all the system are credible.