On the effectiveness of the Meteorological Optical Range for representing visibility as perceived by human drivers and by ADAS employing visible cameras

Abstract

Adverse weather conditions continue to present a significant challenge for Advanced Driver Assistance Systems (ADAS). Quantification of visibility in fog is typically accomplished by measuring the Meteorological Optical Range (MOR). Despite the MOR's extensive recognition and utilisation across diverse domains, including aviation, navigation, and traffic management, there exists a potential discrepancy between the visibility estimated using the MOR, i.e. the optical path length in the atmosphere necessary to diminish the luminous flux of a collimated beam to 5 percent of its original value, and that actually perceived by drivers or camera-based ADAS systems operating within the visible range. Indeed, the anisotropy of the scattering generated by fog particles, in conjunction with the fact that, in the automotive sector, visibility is generally supported by the headlights and street lighting, can lead to phenomena not considered by the MOR. The present study proposes a measurement method and setup for investigating the degree to which visibility estimated exploiting the MOR represents the visibility perceived by drivers or camera-based ADAS systems. The experimental findings indicate that the lighting angle can substantially influence visibility. Thus, the study will propose a simplified analysis to identify the primary contributors to this discrepancy.