Due to the high volume of traffic, many transportation agencies have recognised the advantages of exclusive bus, taxi or carpool lanes and encouraged their funding. But enforcement of the rules of these lanes is very difficult. To mitigate danger for officers performing visual observation, automated monitoring systems have been proposed. Recently, Near-infrared (NIR) camera systems have been investigated for occupancy estimation task. We undertook a study to determine if the automatic NIR detection and counting systems of vehicle passengers are feasible when cars are equipped with a modern solar windshield or heavy tinting. Such automated passenger counting systems would greatly improve the effectiveness of carpool lanes and accuracy to help with the future design of infrastructure. In the present paper, we report our findings regarding the spectral transmittance of different car glasses, car window tintings and solar windshields. We are currently building a novel system based on the fusion of five different image types captured with cameras equipped with NIR, polarising, UV and neutral density filters and we demonstrate the feasibility of the NIR approach with theoretical and experimental arguments.