摘要
足球机器人全局视觉系统为决策系统提供机器人位姿信息,决策系统进行机器人的路径和动作规划;竞赛规则要求视觉系统必须现场安装调试,因此标定过程必须快速有效,且具有较高的精度;针对足球机器人全局视觉系统的特殊应用场景,提出了一种基于两步法简化标定模型的快速有效标定算法;利用近似欧拉角旋转矩阵建立摄像机标定模型,确定需要求取的摄像机内外参数,根据标定模型,对内外参数的求解过程进行了推导;借助场地上现有的一些标志点作为靶标,对模型的内外参数进行分步求取初始值,并利用L-M优化算法对摄像机内外参数进行分步优化求取精确值;通过实验分别从标定误差和机器人动作实现效率两个方面与其他标定算法进行对比,验证了该算法具有较高的标定精度;实验表明,本算法的优点在于不用借助专门的标定靶标,即可快速有效完成相机标定,标定精度可达±1 mm,完全能够满足决策系统的控制精度要求,且标定过程简单易实现。
The global vision system of soccer robots can provide robot posture information for strategy system,which will plan moving path and action for robots.According to match rules,vision system must be installed and tested in field,which requires a fast calibration process with higher precision.For special application in global vision system of soccer robots,a fast and effective calibration algorithm was proposed based on two-stage simplified calibration model.The approximate Euler rotating matrix was used to build the camera calibration model with the intrinsic and extrinsic parameters to be achieved,and the solving process of parameters was derived according to the calibration model.Some mark points on the field were used as reference targets to obtain the initial value of parameters step by step,and the L-M method was adopted to optimize all the parameters step by step.Compared with other calibration methods from two aspects of calibration error and robot movement realization,the proposed algorithm verified the high calibration precision by experiments,respectively.Experimental results show that this algorithm easily and effectively realizes the calibration precision of 1 mm without special reference target,which can completely meet the requirement of control precision for the strategy system,and it is simple and easy for the calibration process to implement.
作者
李奎
陈华
王婷婷
刘波
LI Kui;CHEN Hua;WANG Tingting;LIU Bo(College of Mechnical and Electrical Engineering,Hohai Universtiy,Changzhou 213022,China;College of Science,Hohai Universtiy,Nanjing 210098,China)
出处
《计算机测量与控制》
2024年第5期224-229,共6页
Computer Measurement &Control
基金
江苏省自然科学基金面上项目(BK20201159)。
关键词
足球机器人
全局视觉系统
相机标定
两步法简化标定模型
L-M优化
soccer robot
global vision system
camera calibration
two-stage simplified calibration model
L-M optimization
