전방 카메라를 이용한 주행 가능 영역의 노면 종류 인식 방법

Abstract

This paper proposes a road surface type classification method using front—view camera. The proposed method classifies road surface types in drivable area as asphalt brick and unpaved road. The drivable area is extracted from a front—view image using deep learning. Furthermore, the road area is divided using LiDAR data to deal with the road sections in which road surface type changes. Experiments demonstrate that the proposed method recognized the various types of road surface in the drivable area.