A 3D Vision Pose Recognition Method for Automated Spraying of Segment Molds

Abstract

This invention provides a 3D visual posture recognition method for automatic spraying of tunnel segment molds, belonging to the field of automatic spraying technology for tunnel segment molds. Specifically, it includes: acquiring 2D images and 3D point clouds of the tunnel segment at a standard position A; obtaining a standard feature description D and a standard teaching trajectory Q based on a feature recognition algorithm; after the standard model is established, entering the actual workflow, and acquiring 2D images and 3D point clouds of the tunnel segment at position A′; obtaining the current feature description D′ based on the feature recognition algorithm; calculating the offset M of the tunnel segment at position A′ by comparing the difference between D and D′; superimposing the offset M onto the standard trajectory Q to obtain the offset-corrected spraying trajectory point posture Q′; and having a gantry-driven robotic arm move and spray according to the offset-corrected spraying trajectory point posture Q′ to achieve scanning and spraying of the entire tunnel segment. This invention reduces spraying defects caused by positional deviations and improves spraying efficiency.

Images