Yes, as Lukstr pointed out the moving road and grass is made up of strips of the original image. I think I know how it is done (or at least can be done):

From the bottom of the screen to the middle of the screen there are movie clips placed with decreasing height, say 10 at the bottom and 1 in the middle of the page and everything in between. Every "strip movie clip" is masked using a mask with the same height as the strip movie clip itself (1-10). Inside the strip mc the image is placed and scaled, smaller scale for strips with smaller height, to give the appearence of perspective as Lukstr pointed out. The car has an angle (the direction of movement). The car has a position on the huge image. The images in the strip mc:s are rotated by the car angle around the car position coord on the image. The images in every strip mc are moved to this position, with an offset in y-pos in the strip mc depending on the strip mc's position between the horizon and the bottom of the screen.

The collision detection is probably a comparison to a 2d array (it doesn't have to be that detailed because the road is pretty large, and you don't need that precision...)

PercyPea, have I misunderstood completely? Again great idea and execution of the idea!