Roborealm has an interesting tutorial on using machine vision for obstacle avoidance in ground-based vehicles. The tutorial shows a series of image processing steps which can be chained together to detect the "highest" point in an image, based on the ground color. Presumably, that highest point is indicative of the best path to take in order to avoid the most obstacles.
The ... algorithms will refer to aspects of this images and exploit attributes that are common in obstacle avoidance scenarios. For example, the ground plane assumption states that the robot is placed on a relatively flat ground (i.e. no offroading for these robots!) and that the camera is placed looking relatively straight ahead or slightly down (but not up towards the ceiling).
RoboRealm is a fascinating and powerful tool for exploring machine vision without difficult programming or an in depth knowledge of the field. It is free for personal use and comes with a number of plugins for using it with various robotic kits. If you're curious about machine vision, it's worth a look.