Three-dimensional Positioning Guidance of Automobile Doors Based on Machine Vision

At present, in Workshops, parts fixed on fixtures were commonly guided and gripped by pre-teaching the motion trajeetories of robotic arms, resulting in poor human-machine interaction and low levels of intelligcncc. Machinc vision, as an effective means of non-contact sensing and meas-urement, was applied in industrial scenarios. Based on monocular vision pose measurement technolo-gy, a multi-camera PnP positioning model and robotic arm guiding model were established to achieve spatial pereeption and guided grasping of the parts. Additionally, a pre-calibration method was pro-posed using multi-zero anchors to compensate System errors between robotic arm and vision modulc, which effectively decreasing the positioning and guiding errors of large parts. © 2024 Chinese Mechanical Engineering Society. All rights reserved.