Lane recognition and preceding vehicle tracking based on computer vision

The objective of this research is to develop a vision based intelligent vehicle system (IVS) that can recognize the lane and detect the preceding vehicles. By identifying the lane markings pixels on the urban road, the road triangle is extracted as an effective search area for vehicle detection. Vehicle detection is achieved by first using vehicle shadow feature to define a region of interest (ROI), then the interfere left in the road triangle is pruned. After utilizing the methods, such as histogram equalization, ROI entropy and mean of edge image, the exact vehicle rear box is determined. In vehicle tracking process, the predicted box is verified and updated and some important parameters such as relative distance or velocity, the number and type of the tracked vehicle can be calculated to supply to the driver assistance system to react. The system has been tested under different traffic conditions in Hong Kong urban areas. Testing results demonstrate that our new system illustrates good detection and tracking performance. Our contribution in the work includes road triangle, vehicle shadow highlighting, ROI predicted more robustly.