Detection of road regions is not a trivial problem especially in unstructured and/or off-road domains since traversable regions of these environments do not have common properties unlike urban roads or highways. In this paper a novel unstructured road detection algorithm that can continuously learn the road region is proposed. The algorithm gathers close-range stereovision data and uses this information to estimate the long-range road region. The experiments show that the algorithm gives satisfactory results even under changing light conditions.