A vision-based controller provides translational and rotational control signals to a computer or other input driven device. The controller includes a tracked object, positioned in space and having at least a first reference point and a second reference point. The tracked object is capable of three dimensional rotational and translational movement. At least one imaging device, positioned at a distance from the tracked object, generates an image of the tracked object, at plural succeeding times. A processor unit receives the image, comprised of pixel values, from the imaging device; identifies pixels corresponding to a current center of the tracked object, the first reference point and the second reference point; determines a current dimension (i.e., size or radius) of the tracked object; calculates a translational and rotational displacement of the tracked object based on the above information; and generates control signals in accordance with the transitional and rotational displacement.
展开▼