Robot positioner needs no calibration

Robot positioner needs no calibration

Calibration methods are often difficult and inconvenient in robotic vision systems. When cameras are mounted on moving robotic arms, calibration is even more difficult because systems cannot be recalibrated each time the system moves.

To solve this problem, Bill Yoshimi at the Columbia University Robotics Group (New York, NY) has developed a method that allows robots to be precisely aligned and positioned without the need for calibration. This is done by continuously updating the relationships between the imaging system and robotic actuation system. "By extracting control information directly from the image, the technique does not suffer from the errors associated with fixed calibration methods," says Yoshimi.

A camera is attached to a robot arm so that the camera and robot gripper rotate simultaneously. As the camera system rotates about the gripper`s rotational axis, the circular path traced out by a point-like feature such as a small circle or corner projects onto an elliptical path in image space. Projected feature points are gathered over part of a rotation and fitted to the ellipse. The distance from the rotational axis to the feature point in world space is proportional to the size of the generated ellipse. Once an image ellipse has been determined, moving the robot-camera system recovers the connection between camera and robot.

As the rotational axis gets closer to the feature, the feature`s projected path forms smaller and smaller ellipses. When the rotational axis is directly above the object, the trajectory degenerates from an ellipse to a single point. Bill Yoshimi can be reached at [email protected].