Optical distortion correction considering radial and tangential distortion rates defined by optical design

This paper proposes a high-accuracy distortion correction method for optical imaging distortion mapping and computational image predistortion. Due to the establishment of the predistortion is based on the distortion evaluation in the common optical design software CODE V, it can be used in the broad field of optical imaging system. Two distortion correction methods considering radial and tangential distortion rates which defined by optical design are introduced to perform distortion correction, and the forward and inverse image warping interpolation methods are adopted, separately. Simulations are conducted to verify the effectiveness in correcting complex distortion for freeform off-axis optical system. Furthermore, we implement a distortion correction experiment for an augmented reality head-mounted display that suffers from an asymmetric distortion with the maximum tangential distortion over 7%. For the first method, fast and high-accuracy image warping is achieved, and the RMS error of the approximation predistortion image is less than 1 pixel. Meanwhile, the RMS errors of the corrected imaging points are approximately 1.93 pixels vertically and 2.16 pixels horizontally when using a 5-megapixel camera. The second method considers the additional influence of the actual machining and assembly errors of the optical system, and the RMS errors of 0.89 pixels vertically and 0.85 pixels horizontally are achieved.