摘要 |
PURPOSE: To stably detect a shape and an attitude of an object such as a hand. CONSTITUTION: The object is image-picked up by at least three cameras 1-3. Picture information on at least two cameras are selected by a picture information selection part 4 and are given to a detection part 5. The picture information selection part 4 selects picture information on the two cameras in order from a smaller angle which the normal vector of the plane of the object, which is obtained on detection information of the attitude given from the detection part 5, and the camera shafts of the respective cameras make. Thus, the two cameras optimum for image-picking up the object are always selected. |