Research On Interactive Planning And Simulating For Robot-assisted Binocular Stereo Vision System

Vision-based metrology has been widely used in automobile manufacturing, aerospace industry and other fields for its precise and fast measurement. In an actual inspection task, the relative position between the measuring equipment and the object needs to be changed repeatedly to get the complete data measured target because of the limitation of sensor shooting range. At present it mostly depends on human beings to change the position of the sensor for the vision-based measurement system. This way is obviously slow and the overlap measurement regions will produce more redundant data. So it is low efficiency. With the development of robot technology, the robot-assisted measurement systems begin to emerge in which the robot is a movement actuator. To meet the requirements of the system, we need to plan sensor position, robot path and tracking simulation. On the basis of studying the planning of sensor viewpoint placement and sensor viewpoint path, the paper presents a method of interactive measurement planning and simulation with the measurement unit consist of industrial robot and binocular stereo vision measurement equipment in the software. The main content can be summarized as:1. Robot hand-eye calibration principle and Shiu’s method are studied for solving hand-eye matrix.2.Study on the determination method of single measuring area of robot-assisted binocular stereo vision system. Firstly, the measuring area of binocular stereo vision measurement system is determined by the camera imaging constraints. Then the robot working space is determined according to its working principle. Finally, the measuring area of whole system is determined by combining the previous two.3. An automatic sensor viewpoint planning method is presented which is based on model block and accessibility cone and a viewpoint position selection strategy is provided, which is based on robot angle, shooting rate, coincidence rate and so on. An optimized path is formed by the genetic algorithm.4. Robot trajectory simulation method is studied. The robot joint trajectory is planning by the cubic polynomial interpolation, and animation is simulated with Open GL graphics library.5. A robot-assisted binocular stereo vision interactive measurement planning and simulation system has been developed. The system has good user interactivity and animation simulation. It can help users to do sensor viewpoint placement, sensor viewpoint path planning, and robot path tracking simulation, which could optimize the measurement process and improve the measurement efficiency. The actual measurement results on the door parts verify the effects of measurement planning and simulation system.