Control of camera state based on perspective observation

This paper describes a control law of the camera movement to track the object in such a way that it keeps the projected image of the object at a desired place on the image plane. The camera state which is composed of the position and the attitude is regarded as a point in the manifold S=SO(/sub 3/)/spl times/R/sup 3/. For measuring the difference between the two points in S, A distance is introduced. The set of all the desired position of camera states which satisfy the purpose of control for an object on the camera is a submanifold S/sub d/ of S. The minimum distance /spl rho/*(F, S/sub d/) from a point F/spl isin/S to S/sub d/ is explicitly calculated and is regarded as the distance between a point F and a set S/sub d/. A control strategy for the camera movement to reduce p*(F, S/sub d/) is given based on the measurement of the projected image of the object. A controller for this strategy is given and its convergence is proven.<<ETX>>

[1]  Berthold K. P. Horn Robot vision , 1986, MIT electrical engineering and computer science series.

[2]  B. Ghosh,et al.  A Perspective Theory for Motion and Shape Estimation in Machine Vision , 1995 .

[3]  Lee E. Weiss,et al.  Dynamic sensor-based control of robots with visual feedback , 1987, IEEE Journal on Robotics and Automation.

[4]  Tsutomu Kimoto,et al.  Manipulator control with image-based visual servo , 1991, Proceedings. 1991 IEEE International Conference on Robotics and Automation.