高精度摄像机标定模板的设计及识别算法  

Design and Detection Methods for Accurate Camera Calibration Targets

作　　者：梁力[1] 尹东斐[2] 王川[2] 

机构地区：[1]西安交通大学软件学院,西安710049 [2]西安交通大学电子与信息工程学院,西安710049

出　　处：《西安交通大学学报》2011年第4期82-85,共4页Journal of Xi'an Jiaotong University

摘　　要：针对目前摄像机标定模板——黑白棋盘格识别精度低导致标定精度降低的问题,提出了一种以渐变圆形为特征的摄像机标定模板以及相应的检测算法.新的摄像机标定模板以渐变圆形作为图元代替以前的棋盘格交点,从而可以使摄像机更精确地对其位置进行识别,得到更精确的摄像机标定结果,并保留了黑白棋盘格作为标定模板,具有制作简易,成本低廉的优势,并且针对这种标定模板,提出了精确的检测算法.实验结果表明,在相同的摄像机和相同的环境下,新标定模板和检测算法标定的投影误差比黑白棋盘格的标定结果减少了20%.A new camera calibration target and its detection method are presented to improve the low accuracy of existing common checkerboard detection methods. The new calibration target uses gradient circle as its primitive element, and thus has easy to make, cheap and more accurate features compared with the checkerboard. It is easy and accurate to detect the new designed camera calibration target since the new calibration target contains more information than common checkerboard. A more accurate detection method is provided for the new calibration target. Experimental results and comparisons with the checkboard show that the new target with the detection method reduces the projection error by 20%.

关 键 词：摄像机标定 标定模板 检测算法 

分 类 号：TN391.4[电子电信—物理电子学]