Parameter Identification and Scanning Pose Optimization for Robotic Vision Measurement System

Abstract

Robotic vision measurement system (RVMS) has received wide attention due to their advantages, such as noncontact, flexible inspection, and strong adaptability to various manufacturing environments. However, compared to contacting coordinate measuring machines, larger measurement errors are the main reason that limits the wide application of RVMS in high-accuracy measurement scenarios. To address this problem, a systematic measurement error modeling method and control strategies including parameter identification, error compensation, and scanning pose optimization is proposed to improve the scanning accuracy of the free-form surfaces. At first, multiple error sources in the RVMS are analyzed comprehensively and the influence of kinematic pose errors of the robot, hand–eye pose errors, and sensor geometry structure parameter errors on the measurement system are quantified. Then, a theoretic error model considering multisource errors is established and a method for simultaneous calibration of RVMS parameters is proposed. Afterward, based on the calibrated parameters, a pose optimization method that integrates the RVMS calibration residuals and the measurement uncertainty constraints of the scanning pose is proposed. At last, a real case study was used to illustrate the procedures of the proposed methods. Results showed that the measurement error was significantly reduced compared to the benchmark method.