SCARF - A COLOR-VISION SYSTEM THAT TRACKS ROADS AND INTERSECTIONS

SCARF is a color vision system that recognizes difficult roads and int ersections. It has been integrated into several navigation systems tha t drive a robot vehicle, the Navlab, on a variety of roads in many dif ferent weather conditions. SCARF recognizes roads that have degraded s urfaces and edges with no lane markings in difficult shadow conditions . It also recognizes intersections with or without predictions from th e navigation system. This is the first system that detects intersectio ns in images without a priori knowledge of the intersection shape and location. SCARF uses Bayesian classification, a standard pattern recog nition technique, to determine a road-surface likelihood for each pixe l in a reduced color image. It then evaluates a number of road and int ersection candidates by matching an ideal road-surface likelihood imag e with the results from the Bayesian classification. The best matching candidate is passed to a path-planning system that navigates the robo t vehicle on the road or intersection. This paper describes the SCARF system in detail, presents results on a variety of images, and discuss es the Navlab test runs using SCARF.