VISION-BASED HIGHWAY OVERHEAD STRUCTURE DETECTION SYSTEM

An image sequence is inputted 200 from the camera and vertical motion is estimated 202. A windowed horizontal edge projection 204 is extracted from the inputted image sequence 200 and the horizontal edges are projected 206. The horizontal edge projection 206 and the vertical motion estimation 202 are combined in a horizontal segmentation and tracking element 208, and forwarded to an object parameter estimation element 210 where the object's distance and height are estimated. This data is combined in a fusion with radar detection element 212. By correctly matching the overhead objects sensed by the radar and video camera, the proximity and relative speed can be ascertained. Once overhead objects have been identified they can be isolated and not considered for collision avoidance purposes.