August 15, 2025

Vision-based calibration method for spraying tools

Key Points

Calibration accuracy improved with a maximum error of 3.98 mm in the vision-based method, surpassing traditional techniques.
The study reveals that the vision method reduces calibration time and enhances overall efficiency in robotic spraying applications.
Pose transformation was calculated using a combination of an industrial camera and gradient descent optimization techniques.
Vision-based calibration offers advantages over traditional methods, showcasing a maximum error reduction compared to the four-point method.

Abstract

Aiming at the problem of spray tool (spray gun) calibration of spray robot, a spray tool calibration method based on vision is proposed. Firstly, the data image of the cooperative target installed at the end of the spray robot is collected by an industrial camera. The initial value of the relative pose relationship between the cooperative target and the end of the robot is calculated by the calibration model, and the calibration parameters are optimized by the gradient descent algorithm. Finally, the pose transformation matrix of the spray tool coordinate system is obtained according to the pose relationship between the internal coordinate systems of the cooperative target. Finally, the calibration accuracy is analyzed by the winding point test: the maximum error of the spray tool calibration method based on vision is 3.98 mm, and the average error is 3.01 mm. The maximum error of the traditional four-point method is 4.41 mm, and the average error is 4.02 mm. Compared with the traditional four-point method, the calibration accuracy is improved and the time required for calibration is reduced.

Ask AI

Helpful

Bookmark