| 摘要 |
PROBLEM TO BE SOLVED: To solve the problem in a conventional method that a mobile image detected on the basis of one frame-to-frame difference is different from the shape of a mobile object itself since it includes an image one frame before, and the precise shape or position of the mobile object cannot be detected. SOLUTION: In this detection method of the mobile image, in order to detect the precise shape of the mobile object, binary images A and B are obtained by determining two frame-to-frame differences among three frames and subjecting them to threshold processing. Although the binary images A and B include excessive images except Fr<SB>1</SB>, the mobile object detection result only for Fr<SB>1</SB>can be taken out by determining AND. In this detection method, the position of the mobile object can be detected, in addition to the acquisition of a cutout image of the mobile object, by determining AND of the precise mobile object detection result by the above means and the image one frame before. Further, a voice frame memory is included, whereby a voice with one frame delay synchronized with the detection result can be outputted. COPYRIGHT: (C)2004,JPO
|
