Visual servoing control of 4-DOF palletizing robotic arm for vision based sorting robot system

In this paper, an image-based visual servoing is integrated into a vision-based sorting robot system to increase the flexibility and performance of the system. A novel decoupled image-based visual servoing is proposed to control a 4-DOF robot arm for grasping products. Area, orientation angle, and centroid features extracted from the image are used as input to control the velocities of the robot arm. A multi-threshold algorithm is presented to detect and classify objects and extract the image features. Due to its simplicity but efficiency, image processing and visual servoing algorithms can be implemented in real-time on a low-cost embedded computer. Furthermore, the system is also easy to set up because the robot and camera calibrations are not required. Experiment results reveal the time effectiveness and performance robustness of the system. The system can be used in industrial processes to reduce the required time and improve the performance and flexibility of the production line.