A vision-based wheel-shaped tool calibration method for robot processing system

Robot processing technology is widely used in machining and surface strengthening areas due to the advantages of intelligence and flexibility. Tool calibration is essential for the robot processing system to control the processing trajectory accurately. For components with complex surfaces, the wheel-shaped tool has significant advantages of reaching delicate regions. However, existing tool calibration methods are not suitable for the wheel-shaped tool and face problems such as inaccuracy and complex operation. In this paper, a vision-based wheel-shaped tool calibration method is proposed for robot processing system. Firstly, a novel hand-eye calibration method is proposed to achieve accurate rotation matrix and transfer vector between the vision scanner and the robot base. Then, the wheel annular surface is scanned and extremum points are extracted to calibrate the TCP (tool-center-position). Special tool planes are scanned and their normal vectors represent the TCF (tool-coordinate-frame). Furthermore, the TCP and TCF are used for the hand-tool transformation relationship calibration. Finally, a robot ultrasonic rolling system using a wheel-shaped tool is established, and tool calibration experiments are carried out to verify the feasibility of the proposed method. Additionally, the proposed method is evaluated in repeatability accuracy and calibration error.