摘要
采用基于图象的视觉伺服方法 ,绕过了 3D空间的重建 ,直接利用图象特征来控制机器人的运动 给出了一种基于图象的无标定的视觉伺服控制方法 ,通过动态估算图象雅可比矩阵来代替实际的雅可比矩阵 ,消除了工业应用中标定的繁琐和由此引起的误差 。
In a position based visual servo control system,a serial of properties extracted from image are used to estimate the relative position between target and camera.The nonlinear imaging process makes the visual reconstruction difficult.An image-based visual servo that bypassed 3D space reconstruction and made use of image properties to control robotic motion.Image Jacobian provides a crucial contribution in the image based visual servoing control.To work out the robotic end effecter motion to targer,it is necessary to calibrate robot, camera and the relation between them.The image based visual calibraton-free servo control dynamically estimates the image Jacobian to replace actual image Jacobian. This method eliminated the tedious and error prone calibration in industrial applications,and especially avoided the difficulty of visual reconstruction that brought about by the nonlinear imaging process and image matching in the position based visual servoing control method.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2002年第3期294-296,共3页
Journal of Harbin Institute of Technology
基金
国家自然科学基金重点资助项目 (5 963 5 160 )