ML Models for Winter Road Surface Condition Recognition in the Case of Insufficient Coverage

To maintain roads during the winter season, it is critical to determine their surface conditions, for efficient allocation of cleaning efforts such as plowing or salting. Current techniques for determining road surface conditions include the use of expensive roadside weather stations or sending trained personnel for on-site inspection. However, these stations are only placed on some roads, and they have a limited coverage. This paper proposes a method to determine road surface conditions in places with insufficient coverage.