| 摘要 |
The method of processing a sequence of images representing a road scene taken by a camera installed on board a vehicle for extracting the road and detecting obstacles consists, on the one hand, of a stage (2) of static analysis of the scene from a CCD-type sensor (1), in combining an analysis of the luminance information from the video signals with an analysis of the colour video signals of an image, and consists, on the other hand, from the same sensor (1), in a stage (3) of dynamic analysis of the same scene, in validating the information arising from the stage (2) of static analysis of the scene, and in monitoring the movement of the obstacles assessed as dangerous, all through the sequence of images, according to a criterion which is defined with respect to the vehicle. It makes it possible to present the user of the vehicle with rapidly usable information so that he can, if appropriate, change the forward speed of the vehicle on the basis of the significance of the danger. This method is applicable particularly to road safety and to mobile robotics. <IMAGE>
|
