Abstract—This paper introduces a new method which normalizes spectral profile in hyperspectral images. In order to use Visual Light Communication, it should be detected some cars in front of a driver. However, there are some problem in methods using CCD camera for the detection of vehicle. CCD cameras regard the streetlamps as automotive headlamps. In order to solve this problem, hyperspectral camera can be a solution. Because hyperspectral images have hundreds of bands, information of hyperspectral images more than the CCD images. However it is necessary to preprocess hyperspectral image to remove noise and illumination dependency. This paper proposes more effective spectral normalization using Per-Norm. Also it show real experimental results demonstrate the robustness of the proposed method.

Index Terms—Automotive headlamp, hyperspectral image, LED spectral, per-norm, spectral normalization, visible light communication.

The authors are with the Department of Electronic Engineering, Yeungnam University, Gyungsan-si, Gyeongbuk 712-449, Korea.

Cite:Heekang Kim, Sungho Kim, "Per-Norm Based Automotive Headlamp Detection in Hypersepctral Image for Visible Light Communication," International Journal of Computer and Electrical Engineering vol. 8, no. 1, pp. 12-21, 2016.