一种角点自动检测与排序算法

An Algorithm of Chessboard Corners Automatically Detection and Sorting

为了提高摄像机标定精度,通过分析当下比较流行的几种角点检测算法的优缺点,结合Hessian矩阵理论,首先通过动态自适应确定角点和剔除伪角点,然后应用最小凸包原理自动识别拐点,对角点进行排序和亚像素定位。通过不同图样多次实验证明,该算法适用范围广、精度高、鲁棒性好。

To improve the accuracy of camera calibration, the advantages and disadvantages of several popular contem- porary corner detection algorithm are analyzed. Then combined with Hessian matrix theory, firstly the corner points are dy- namic self-adaptive determined while false corners are eliminated, secondly the turn points automatically recognizing and the chessboard corners sorting and sub-pixel locating are accomplished by using the Convex Hull principle. It has been proved by experiments to be an algorithm with high precision and good robustness.