Road boundary detection for autonomous navigation

As image processing continues to mature, its areas of application continue to expand. Autonomous land navigation is one arena that has benefited from its advances in recent years. Autonomous land navigation is a process of piloting a vehicle from one location to another desired location. There are two major components of navigation, obstacle avoidance and 'boundary' detection for the path to be traversed. This work centers on the 'boundary' identification aspect of navigation, specifically, it deals in the automatic detection of road boundaries with a monochrome image of the visual scene. The approach presented employs the statistical information of the gray levels present in an image coupled with constraints based on the road geometry. >