Welding Seam Tracking And Correcting System Based On Laser Vision

This dissertation aimed at the low accuracy of teaching welding robot and the robot cannot correct the welding position in real-time; researched on the development of welding robot at home and abroad; found that welding robot technology is bounded together with sensor technology, it showed that laser vision achieves the largest information of welding seam and has high accuracy among those sensors which can be used on welding seam tracking and correcting. Thus, this thesis proposed a method that added a laser vision module on the platform of teaching robot to improve its accuracy. All of the work including:(1) Designed the hardware and software of the laser vision system. The design and selection for hardware equipments include: industrial camera, optical lens, laser sensor, and narrow-band filters, in addition, this dissertation designed a protection device for the laser vision system.(2) The software of this system was developed on the environment of VS2008 and Open CV, the main work of image processing can be divided into two steps: preprocessing and postprocessing, the purpose of preprocessing is reduce image noise and enhance the image information of optical stripe, methods such as Gaussian filtering, histogram equalization and image binarization.(3) After image preprocessing, the clear optical stripe can be obtained, then steped into image postprocessing. In this part, the region of interest was chosen first to narrow the image processing range and reduced time for subsequent process; then, through image thinning method to get the center line of optical stripe; at last, used slope method to extract feature points from image, also analyzed extraction results and image errors calculation.(4) Analyzed the camera calibration principle, and through the calibration to determin the conversion relationship between image coordinate system and world coordinate system; acquired calibration error after calculation; established RS232 serial communication between PC and teaching welding robot controller TMS320F2812.Combined camera calibration error with image error, the total error of vision system is less than 1 mm, this system has improved the accuracy of teaching welding robot.