| 摘要 |
PURPOSE:To recognize an obstacle on a drive path by extracting an edge of an entire picture and a drive lane mark edge and subtracting the drive lane mark edge from the picture edge. CONSTITUTION:A 1-frame video signal of a camera 4 is ADD-converted and stored in a frame memory 3. In order to arrange steep leading and trailing edges in picture data, smoothing by spatial filtering is made. Then the spatial filtering is adopted to extract a drive lane mark edge. The drive lane mark edge is extracted separately for an inner lane mark and an outer lane mark. An obstacle on a drive path is recognized by a picture processing unit 2 by subtracting the drive lane mark edge from the picture edge extracted in this way. Thus, the obstacle is recognized from a road picture without need for complicated processing. |
