CAJ | 학술논문

介绍了单目视觉系统标定的现状，以及现行标定法的优缺点。在相机固定、机器人末端执行器在平面内运动的场合，提出了一种基于最小二乘法的拟合标定算法，该算法可以考虑径向畸变、切向畸变等非线性因素。在不同的拟合模型下，该标定算法分别将平均标定误差从原来的1.4505mm下降到0.4879mm、0.1084mm，最大标定误差从原来的2.2001mm下降到0.7587mm、0.2081mm，大大降低了标定的误差。并进行了算法的可靠性验证，证明了算法的有效性。
개소료단목시각계통표정적현상，이급현행표정법적우결점。재상궤고정、궤기인말단집행기재평면내운동적장합，제출료일충기우최소이승법적의합표정산법，해산법가이고필경향기변、절향기변등비선성인소。재불동적의합모형하，해표정산법분별장평균표정오차종원래적1.4505mm하강도0.4879mm、0.1084mm，최대표정오차종원래적2.2001mm하강도0.7587mm、0.2081mm，대대강저료표정적오차。병진행료산법적가고성험증，증명료산법적유효성。
The research status of single camera vision system calibration and current calibration methods’ advantages and disadvantages is introduced. In the situation where camera is fixed and robot end effector moves in a plane, a camera calibration method based on least squares method is proposed, considering the nonlinear factors of radial distortion, tangential distortion, etc.. In different fitting models, the average calibration error is reduced from the original 1.4505mm to 0.4879mm, 0.1084mm respectively, and the maximum calibration error is reduced from the original 2.2001mm to 0.7587mm, 0.2081mm. This method greatly reduces the camera calibration error. The reliability verification experiment is also conducted to prove the effectiveness of the method.