A Robust Camera-Based Method for Optical Distortion Calibration of Head-Mounted Displays

One of the problems in using HMDs is that the virtual images shown through the HMDs are usually warped due to their optical distortions. In order to correctly compensate the optical distortions through a predistortion technique, accurate values of the distortion parameters are required. Although several distortion calibration methods have been developed in prior work, these methods have some limitations. In this paper, we propose a method for accurately estimating the optical distortion parameters of both immersive and (optical) see-through HMDs, without the limitations. The proposed method is based on photogrammetry and considers not only the radial distortion but also the tangential distortion. Its effectiveness, including the effects of different coefficient orders of radial and tangential distortions, are evaluated through an experiment conducted with two different HMDs, an optical see-through head-mounted projection display (HMPD) and a commercially available immersive HMD. According to the experimental results, the proposed method showed significantly lower reprojection and line-fitting errors than a previous method proposed by Owen et al., and the radial distortion coefficients estimated by the proposed method were significantly more accurate than the nominal values obtained from the HMD optics prescription. In addition, when both the radial and tangential distortions were considered, the proposed method was significantly more effective than when only the radial distortion was considered. We also discuss an OpenCV-based simple implementation of a distortion correction method.