Pollution Discrimination on Rail Surface for Adhesion Evaluation Using Hyperspectral Signatures
Résumé
The ability to evaluate wheel-rail adhesion of trains on rail tracks is important in various fields of Intelligent Transportation System ITS: Environment Perception, Safety and Driver Support, Automatic Driving, Transport Management. Adhesion is degraded when rails are polluted. Each pollution material has its proper impact on adhesion. Moreover, each material (due to its composition) reflects the light in its own way, allowing it to differentiate itself from others. In this paper we will characterise recurrent rail pollutant using a custom light and spectrometer system and associate each spectrum with an adhesion or friction coefficient.