With the continuous development of manufacturing industries both domestically and internationally, there is an increasing demand for the 3D reconstruction of large-scale workpieces. In order to expand the measurement range while ensuring measurement accuracy, a 3D reconstruction method based on multi-camera wide-field tracking is proposed. First, calibrate the relative poses of the multi-camera system and establish the poses of all camera coordinate systems with respect to the global coordinate system. Secondly, attach circular markers around the line laser sensor and establish a marker coordinate system, calculating the relative pose between the marker coordinate system and the line laser sensor coordinate system. Thirdly, determine the camera field where the line laser sensor is located, and obtain the pose of circular markers using a pose matching algorithm based on geometric distance. Finally, combining the pose of circular markers and the relative poses of cameras, unify the locally acquired data from the line laser into the global coordinate system, achieving data stitching. The experimental results indicate that within a range of 1.1 meters, the average translation error of this method does not exceed 0.38mm, and the average angle error does not exceed 0.31°, demonstrating the capability for 3D reconstruction of large-scale workpieces.
|