Comprehensive Improvement of Binocular Structured Light Calibration Method Based on Radical-Conservative Cooperative Particle Swarm.

To achieve high-precision 3D reconstruction, a comprehensive improvement has been made to the binocular structured light calibration method. During the calibration process, the calibration object's imaging quality and the camera parameters' nonlinear optimization effect directly affect the caibration accuracy. Firstly, to address the issue of poor imaging quality of the calibration object under tilted conditions, a pixel-level adaptive fill light method was designed using the programmable light intensity feature of the structured light projector, allowing the calibration object to receive uniform lighting and thus improve the quality of the captured images. Then, collaborative Particle Swarm Optimization was studied to optimize the camera parameters. Compared with other optimization algorithms, this algorithm has higher global search capability and can obtain more accurate camera parameters. Under comprehensive improvement, the 3D reconstruction accuracy of binocular structured light is 0.053 mm, showing a 36.33% improvement in reconstruction accuracy compared to mainstream calibration methods.