The Robot Welding Position Locating System Based On Structured-light

This dissertation is studied based on the application of the small special welding robot for containers. The structured-light image processing of welding seam and the location method of robot vision system is studied in depth. What is the most important for robot auto-welding is to acquire the accurate information of the initial welding position and send them to the robot controller. According to the characteristic of reinforced plate welding and the container production environment, a simple and practical locating system of initial welding position is designed for reinforced plate in this dissertation.Firstly, this dissertation summarizes the robot vision system and structured- light measuring technique, also describes their development, research status and the application in welding briefly. It focuses on the methods of robot vision system location and linear structured-light image processing and also summed up the common methods of the extraction of structured-light stripe center and edge measurement. Based on this, the dissertation presents the algorithms for initial welding position locating system. Its vision sensor system of which is comprised of two lasers and one CCD camera, the hardware and software of the system are also designed.Because of the coarseness of the container welding environment and the characteristics of normal size, fixed welding route and small welding area, this dissertation uses the two location methods which are based on fixed in workspace vision system with linear structured-light and a new method based on eye-in-hand vision system. And the methods of welding seam feature recognition are used that based on center of structured-light stripe and based on edge of structured-light stripe respectively.The above algorithms are tested on MOTOMAN-SV3 robot test platform. The two different structure vision locating systems is analyzed and compared from the data results and error of experiments. This dissertation also presents the processing results of the welding seam structured-light image and the location of initial welding position. The experiment result shows that the method presented in this dissertation is practical and reliable and the precision meets the actual needs of production.