라이다와 시공간 정보를 활용한 도로 노면 상태 판단 알고리즘

Abstract

In this paper, we propose a classification method of road surface conditions using LiDAR with spatio-temporal information. We utilize the reflectivity information available through the LiDAR sensor and can have robust properties. Feature vector is selected the reflectivity and the total number of data points of the LiDAR. We propose the spatio-temporal structure for the Deep Neural Network (DNN). The road ahead of the vehicle is divided into four regions to utilize spatio-temporal information. And stack the data to use the time-windowing method. The final results can be obtained by utilizing geometric information from the result of DNN in each region.