A Warning System for Obstacle Detection at Vehicle Lateral Blind Spot Area

For a real driver assistance system, the weather, driving speed, and background could affect the accuracy of obstacle detection. In the past, only a few studies covered all the different weather conditions and almost none of them had paid attention to the safety at vehicle lateral blind spot area. So, this paper proposes a hybrid scheme for pedestrian and vehicle detection, and develop a warning system dedicated for lateral blind spot area under different weather conditions and driving speeds. More specifically, the HOG and SVM methods are used for pedestrian detection. The image subtraction, edge detection and tire detection are applied for vehicle detection. Experimental results also show that the proposed system can efficiently detect pedestrian and vehicle under several scenarios.