Knowledge-Based Road Extraction from High Resolution Remotely Sensed Imagery

Road tracking is a promising technique to increase the efficiency of road mapping. In this paper, a new road tracker, angular texture signature, is proposed based on the knowledge of the roads on high resolution imagery. The tracker uses parabola to model the road trajectory and predict the position of next road centreline point. It employs angular texture signature to get the moving direction of current road centreline point, and move forward one preset step along the direction to reach a new position, and then use curvature change of the parabola to verify the new added road point whether right enough. We also build compactness of angular texture signature polygon to check whether the angular texture signature is suitable for tracking. Extensive experiments demonstrate that our tracker is robust enough to be used to track road on high resolution imagery.