摘要
针对基于视觉测量的物体位姿测量系统要求标定系统精度较高的需求,提出了一种自动识别和亚像素提取黑白平面棋盘格模板图像内部角点的方法.该方法在详细分析了棋盘格图像局部特性的基础上,建立了准确高效的改进SUSAN角点检测算法和精确的亚像素级提取角点坐标的方法,计算出总体标定误差为0.2个像素.通过实验验证了该算法的可行性和有效性,能够为高精度标定系统提供可靠数据.
In order to meet the high-accuracy demand of the camera calibration of the object pose measurement system,an automatic recognition and sub-pixel positioning interior corner coordinate method of the plane white-black chessboard image was proposed.The method established an improved SUSAN algorithm that is accurate and efficient based on the detailed analysis of the local characteris-tics of the chessboard image,introduced the accurate way to position the sub-pixel corner coordinate and then showed the calculated total calibration error is 0.2 pixels.At last the feasibility and effec-tiveness of this algorithm were verified by the experiment.The proposed algorithm can supply camera calibration system with reliable data.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2015年第S1期170-172 177,177,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
大连民族大学博士科研启动基金资助项目
关键词
图像处理
棋盘格图像
角点检测
亚像素提取
提取方法
image processing
chessboard image
corner detection
sub-pixel positioning
processing method
