Design And Research On Vision Sorting And Tracking System Of Delta Parallel Robot

This paper presents a robot-based sorting and tracking system using industrial computer and embedded vision, in which a Delta parallel robot, industrial control computer, Embedded development board, industrial cameras and a belt-based conveyor. The operation efficiency of the system are greatly improved through embedded vision technology. The proposed system can be applied in induatrial products sorting and tracking. The works of this paper mainly includes:1) Calibration between a robot,vision systems and a belt-based conveyorIn view of the complexities of the calibration between the double cameras and multiple systems, the first step is to complete the calibrations of kinematics of the robot and cameras are implemented firstly, and, Throung these calibrating information, both the calibration between the conveyor belt and vision system and the calibration between the belt-based conveyor and the robot are implemented by using the calibrated cameras, the robot, the laser tracker, Matlab calibration toolbox. These calibraitons make it possible for precision manipulation.2) Algorithm research on visual detection and localization of the objectsThe connected zones of the binary images and their geometric centers can be quickly extracted by a improved run-length-coding method by adding label tables enhancing its finding capability. And then the minimum enclosing rectangle can be exacted rapidly using the MER algorithm, which can be used to compute the position and posture of the target objects. The equations of the boundary straight lines are extracted by the new presented algorithms of a binary image searching and boundary line detection using he extracted boundary points of the connected zones.. And the position and posture of the tobjects are obtained according to the fundamental characteristics of intersecting point between the lines and other informations from another vision system. Finally the real-time image acquisition and processing systems based on two cameras are developed by using VS and QT integrated development environment respectively.3) Robot trajectory planning, real-time tracking and sortingThe robot trajectory planning is implemented according to the extracted position and posture information. The buffer time is presented to make the changes of the motion acceleration more smooth and this suppresss the vibration caused by the sudden change of the acceleration. Small even time-intervals interpolation method is adopted to make the sorting and tracking more accurate and stable.