Meisei University, Graduate School of Science and Engineering, 2-1-1, Hodokubo, Hino, Tokyo, Japan 191-8506.

---

Abstract—We propose a method for detecting the lit portion of brake lamps from rear images of vehicles. Its effectiveness was confirmed for 526 rear images. We used 526 Caltech Cars 2001 (Rear) images published by the California Institute of Technology as input images. These images were photographed in the daytime by the in-vehicle camera. The rear image of the car in front was captured. When the brake lamps were lit, the color was red, and the brightness was high. By determining this color from the image, the lighting of the brake lamps could be detected.

Index Terms—Brake lamp detection, autonomous vehicle, driving assistance, computer vision, image processing.

[PDF]

Cite: Yuki Omori, Yoshihiro Shima, "HSV Color Space Based Lighting Detection for Brake Lamps of Daytime Vehicle Images," Journal of Computers vol. 14, no. 1, pp. 25-30, 2019.