Contrast ratio of road markings in Poland - evaluation for machine vision applications based on naturalistic driving study / Tomasz E. Burghardt, Anton Pashkevich // W: Advances in Road Infrastructure ... więcej

Contrast ratio of road markings in Poland - evaluation for machine vision applications based on naturalistic driving study / Tomasz E. Burghardt, Anton Pashkevich // W: Advances in Road Infrastructure and Mobility : Proceedings of the 18th International Road Federation World Meeting & Exhibition, Dubai 2021 [Dokument elektroniczny] / eds. Amin Akhnoukh, Kamil Kaloush, Magid Elabyad, Brendan Halleman, Nihal Erian, Samuel Enmon II, Cherylyn Henry. – Cham : Springer, 2022. – (Sustainable Civil Infrastructures, ISSN 2366-3405). – S. 676-690 : fot., tab., wykr. – Tryb dostępu: https://link.springer.com/chapter/10.1007/978-3-030-79801-7_49. – Abstr. – doi: 10.1007/978-3-030-79801-7_49. – ISBN 978-3-030-79801-7 (eBook). – ISBN 978-3-030-79800-0 (Softcover)

Road markings are critical road safety features for both human drivers and for machine vision technology used in advanced driver assistance systems and in the emerging technology of automated vehicles. ... więcej

Road markings are critical road safety features for both human drivers and for machine vision technology used in advanced driver assistance systems and in the emerging technology of automated vehicles. Amongst the parameters, contrast ratio is necessary for appropriate recognition of road markings. To assess the contrast ratio of road markings at various roads, still images of roadway were obtained from a dashboard camcorder used for a naturalistic driving study in Poland. Road markings and neighbouring roadway surface were measured for luminance and Weber contrast was calculated. At the studied representative roads average contrast ratio was 0.8 under daytime illumination and 2.0 at night; enhancement of the contrast through digital image manipulation resulted in increases to 2.3 and 6.8, respectively. Under poor visibility daytime conditions (interference from glare or rain), average contrast ratio dropped to 0.5 (enhanced 1.4); in the worst case it was below 0.1. Consequently, the current machine vision technology could fail under some poor visibility circumstances. The image enhancement indicated that both the initial and digitally enhanced contrast ratios were important.