Night time vehicle detection using rear-lamp intensity

Conventional vehicle detection algorithms with rear-lamp mostly utilize color data. However, due to the specification of night-time environment with low-light, the color data often has not fully information. To resolve that problem, this paper contains vehicle detection algorithm using only gray value of rear-lamp in night-time. Unlike implement binarization with the color data and gray value, this paper implement tone-mapping process with only intensity of illumination to preserve the shape of rear-lamp. Misdetection rate increases because of using only gray value, but this can be solved by applying condition of rear-lamp characteristic to proposed algorithm. In addition, misdetection is removed through the tracking algorithm. This process can be utilized in automatic driving system, diriving assistance system and so on that recently has been issued because the proposed algorithm works in real-time.