摘要 |
<p>Instead of comparing pixel values at a corresponding point, a corresponding point is searched according to an error of a ray vector projected on a projection plane. There is no need of using cameras having an identical camera lens and an identical distortion parameter. It is possible to connect picked up images of different camera models. Without converting the picked up image into a pin-hole image, according to the ray vector error, the original picked up image is directly attached to an output frame, thereby suppressing deterioration of the pixels. Accordingly, it is possible to preferably combine various picked up images of cameras having different lens distortion and camera models.</p> |