| 摘要 |
PROBLEM TO BE SOLVED: To provide a picture processor which detects the position and attitude of each of piled objects (workpieces) in the same shape. SOLUTION: Teaching models are generated from two-dimensional pictures obtained by image picking up workpieces by a camera from plural directions, and stored, and the relative position and attitudes of the workpieces with respect to the camera are stored correspondingly to the teaching models. The piled workpieces are image picked up by the camera so that two-dimensional pictures can be obtained. The position and attitudes of the camera at that time are stored (200-202). Then, matching processing is operated between the teaching models and the image picked-up two-dimensional pictures so that the conformed teaching model can be obtained (203-207). Then, the three-dimensional position and attitudes of the workpieces are obtained from the relative position and attitudes of the workpieces of the obtained teaching model with respect to the camera and the position and attitudes of the camera at the time of photographing (208 and 209). The picking of the workpieces is operated by a robot based on the position and attitudes (210). Thus, it is possible to automatically operate the picking of even the piled workpieces. |
