| 摘要 |
The system and the method are intended to enable robots, devices, tools, etc., to see the environment in which they operate, by means of a pair of identical cameras, aligned, compatible and coordinated with their fields of sight (photographing) being parallel and adju sted in parallel divergence (M) between each and every line in the fields of sight, and to identify anything viewed by the cameras immediately and at the rate of photographing/filming. By means of the system and the method, the computer vision receives the pictures from the cameras into a designated place for this purpose and a backup is created for the field of sight from the pictures stored in the spatial-memory, and it calculates the distance for each and every point in the picture, as well as the dimensions of the shapes and registers their various features. The system includes a memory register for movement identification at time intervals and calculation of movement, motion, speed and direction of each and every shape in the pictures of the leading camera, received after color filtering, a register of basic shapes and data table(s) such as the "true" table. Said registers are meant for purpose of detecting data, features, definitions and for drawing of conclusions for the key elements. The data, the features, definitions and conclusions enable to compose keys for the unidentified shapes, compatible with the recognized stored shapes register, and thus identification of the unidentified shapes is carried out fully and almost immediately. |
