Three-Dimensional Localization Method of Multiple Infrared Cameras Based on Subpixel

This paper presents a novel three-dimensional localization method using multiple infrared cameras based on subpixel localization to overcome the low accuracy problem of traditional indoor space localization techniques. In this method, subpixel localization algorithms with the Harris operator are employed during camera calibration and localization experiments to replace traditional pixel localization algorithms in the two-dimensional image positioning stage. The Bundle Adjustment method is used to optimize the camera extrinsic parameters, while the collinearity equation and the least square method are applied to solve the spatial point coordinates. The feasibility of the proposed method is verified through experimental methods. The results show that the three-dimensional positioning system achieves millimeter-scale accuracy, surpassing that of pixel-level positioning algorithms.