| 摘要 |
PURPOSE:To decrease the recognizing processing time by recognizing one object at first and then utilizing the result of position recognition of the first object for the succeeding recognition of objects in recognizing plural objects having a relation of connection. CONSTITUTION:A TV camera is moved by (X1+lxB) in (x) direction and by (y1+lyB) in (y) direction from a center O1 of a reference coordinate by using the relative relation of position between corners PA1 and PB1 as x1=Lx1+DELTAx1 and y1=Ly1+DELTAy1. When the accuracy of movement is + or -DELTAxb toward (x) direction and + or -DELTAyb toward (y) direction, a position of the TV camera at a corner PB1 on an image pickup plane has only to be shifted by + or -(DELTAxa+DELTAxb) toward (x) direction and + or -(DELTAya+DELTAyb) toward (y) direction at maximum to a center O2 of the image picture plane of the TV camera in moving the TV camera to recognize the corner PB1, and the retrieved range to recognize the corner PB1 is limited to a very small range.
|
