Traffic light classification is an important feature for automated driving, and it requires correct separation of the color signals. The authors Korbinian Weikl, Damien Schroeder, and Walter Stechele propose a suitable color distance metric and methods for optimization of the color separation. 

Their paper entitled "Optimization of automotive color filter arrays for traffic light color separation" has been accepted for presentation at the Color and Imaging Conference, CIC 2020 virtual, November 4-19, 2020.