Robust Camera Calibration by Optimal Localization of Spatial Control Points

This paper proposes a novel method for localization optimization of control points for robust calibration of a pinhole model camera. Instead of performing accurate subpixel control point detection by specialized operators, which are normally adopted in conventional work, our proposed method concentrates on estimating the optimal control points in regions of plausibility determined by distortion bias from perspective distortion, lens distortion, and localization bias from out-of-focus blurring. With this method, the two main strict preconditions for camera calibration in conventional work are relieved. The first one is that the input images for calibration are assumed to be well focused and the second one is that the individual control point needs to be detected with high accuracy. In our work, we formulate the accurate determination of control points' localization as an optimization process. This optimization process takes determined control points' uncertainty area as input. A global searching algorithm combined with Levenberg-Marquardt optimization algorithm is introduced for searching the optimal control points and refining camera parameters. Experimental results show that the proposed method achieves higher accuracy than the conventional methods.