An Estimation Method of Visibility Level Based on Multiple Models Using In-vehicle Camera Videos under Nighttime

A method for estimating visibility level based on multiple models is proposed in this paper. In winter roads under nighttime, visibility is affected by snow and road lighting. It is possible to encounter poor visibility and cause traffic accidents. In order to reduce these risks, road administrators and users need to acquire visibility information to avoid roads under poor visibility. Therefore, we estimate a visibility level under nighttime. In the proposed method, we estimate a visibility level based on multiple models. Multiple models of image recognition are composed of machine learning and a mathematical model. By utilizing them cooperatively, it is expected to estimate a visibility level accurately. The effectiveness of the proposed method is verified by actual data.