Detection and tracking of overtaking vehicle in Blind Spot area at night time

The BSD (Blind Spot Detection) is one of the key functions of ADAS (Advanced Driver Assistance System). In this paper, we propose an algorithm for detection and tracking a vehicle in blind spot area at nighttime. Since the appearance-based approach that can be used for BSD at daytime is not any more appropriate, a new feature of blobs that is due to the headlights of behind vehicle was adopted. In addition, we calculated the motion vectors of detected blobs and made headlight pairing by use of the projection map. We can successfully generate an alarm signal by tracking the vehicle in blind spot.