复杂背景下改进的棋盘格内角点检测算法

Improved checkerboard grid corner detection method under complex background

针对当前棋盘格内角点检测算法在复杂背景下表现不佳的问题,提出一种改进的棋盘格内角点检测算法。在分析现有算法的基础上,将自适应阈值的BW算子和环形模板算法相结合,提取内角点的像素级坐标,通过灰度梯度法提取亚像素角点坐标。在MATLAB中进行实验,实验结果表明,在复杂背景下,该算法可以有效检测出棋盘格内角点,定位精度达0.1005pix,能够为标定摄像机提供高精度数据,具有一定的实用价值。

Aiming at the problem that the existing corner detection algorithm in the checkerboard performs poorly under complex background,an improved algorithm for detecting the corner points in the checkerboard was proposed.Based on the analysis of existing algorithms,the adaptive threshold BW operator and the ring template algorithm were combined to extract the pixel-level coordinates of the inner corner points,and the sub-pixel corner positioning was performed using the gray gradient method.Experimental results in MATLAB show that the algorithm can not only effectively extract the inner corners of the checkerboard,but also achieve a positioning accuracy of 0.1005 pix.It can provide reliable data for calibration high-precision cameras and has certain practical value.