Spectral reflectance characterization of the road environment to optimize the choice of autonomous vehicle sensors

Michèle Colomb, Pierre Duthon, Frédéric Bernardin

Abstract: A wide range of sensors is being developed for autonomous vehicles in order to ensure safe driving conditions by detecting any relevant objects in the road environment. Sensors, cameras, Lidars and radars, cover a wide range of wavelength, ranging from visible, Near Infrared, Short Wave Infrared and beyond for high frequency radars. For Lidar technology, a single wavelength scans the environment. To obtain most efficient detection, the choice of the wavelength must be optimized. Knowledge of the spectral reflectance of road objects at the chosen wavelength is a key point. An original application of reflectance measurement techniques is proposed to characterize road objects in the wavelength range 350 to 2450nm. This investigation could lead to the definition of reference tests including physical characterization of the environment.