| 摘要 |
PURPOSE:To improve a processing efficiency by selecting only feature points, and performing an observation by a projection method, after required number of the feature points have been extracted. CONSTITUTION:An object 10 is image-picked up with a television camera 1, and the picture of the object is generated on the image pickup plane of an area sensor 9, and a picture processor 7 selects the first feature point, and designates an image point Q on the area sensor 9. Following that, when a laser beam 11 is outputted, the laser beam 11 arrives at an object point corresponding to the image point Q, then generating an optical point P, and the optical point P is image-formed at the optical position detector 13 of a PSD camera 5, and a coordinate where an image point W is located, is outputted to the picture processor 7 as an analog quantity. Next, at the picture processor 7, the spatial position coordinate of the optical point P is calculated based on a principle of triangulation. By executing an operation repeating the same process, the spatial position coordinate of each point on the object point 10 corresponding to all of the feature points can be found, thereby the recognition of the object being obtained.
|
