针对棋盘格角点的检测,提出了一种运算速度快、定位精度高的Hessian矩阵子像素角点检测算法。 简述了Hessian矩阵的原理、算法的执行流程和黑白棋盘格子像素坐标提取的基本方法。通过实验检 测出了黑白棋盘格中的角点,得到了角点坐标,并利用二阶泰勒展开式得到了角点的子像素坐标。最 后,在不同的噪声水平下,与Harris算法作比较,计算出提取误差与提取时间。实验证明,Hessian矩阵 对黑白棋盘格角点检测得更加精确。该实验结果可应用于摄像机标定中。

For corner detection of black white checkerboard, a new sub-pixel operator with high speed and high position accuracy was proposed. The principle of Hessian matrix, the execution procedure of the algorithm and basic ways of getting the sub-pixel coordinate in black white checkerboard were presented. All the corners needed in black white checkerboard were detected and all the coordinates of corners were calculated through experiments. Sub-pixel coordinates of corners were calculated with two order Taylor expansion. Finally, under different noise levels, compared with the Harris algorithm, extraction error and extraction time are calculated. The experimental results show that Hessian matrix is more accurate for corner detection of black white checkerboard. The result can be used in camera calibration.