Sub-pixels corner detection for camera calibration

Two simple and efficient algorithms of sub-pixel corner detection for camera calibration by using plane pattern were presented. The pixel level corners were detected by classical Harris corner detector, and then some methods were used to get sub-pixel accuracy. In one algorithm, the sub-pixel location was achieved through a quadratic approximation of the 3 × 3 neighborhood of the local maxima of the corner response function, and a closed form solution developed to the problem. In the other algorithm, the sub-pixel accurate corner locator was obtained from the observation that any vector from true corner location to its neighborhood is orthogonal to the image gradient. So the sub-pixel location was calculated iteratively by minimizing an error function. The result of the camera calibration shows that high precision can be acquired by using the sub-pixel algorithms and the mean reprojection error is below 0.15 pixels.