Camera Calibration Algorithm for Long Distance Binocular Measurement

Binocular vision has important applications in long distance measurement fields such as aircraft surface contour reconstruction. To improve the measurement accuracy of binocular vision system under long distance condition, a new binocular camera calibration method is proposed in this paper. The infrared LED point light source with constant brightness is fixed on the cantilever of the CMM(coordinate measuring machine). Through the movement of the mechanical mechanism of the CMM, the LED point light source is accurately moved to the fixed position and the two cameras are used to collect images. Through the spatial movement of LED point light source, a virtual lattice composed of discrete points is formed. Because the distance between two points does not change due to the selection of coordinate system, an optimization function with binocular camera parameters as independent variables is designed and each camera parameter is determined by high-dimensional space function optimization. The experimental results show that without considering the distortion factors, the measurement error of the camera parameters obtained by the algorithm in this paper is much higher than that of Zhang Zhengyou algorithm. In the field of view with the axial depth of 2.5 m-4.5 m, the single point measurement error is less than 2.4 mm.