MoueYuusuke

Investigation of road snow depth movement measurement method using in-vehicle LiDAR

TakahasiKazuyosi

In a snowfall area, snow removal in winter is an indispensable task for stable social life. In recent years, however, it has become difficult to maintain a snow removal system due to the effects of public budget cuts and aging. In order to maintain the snow removal system in the future, it is necessary to reduce the cost of the entire work and to efficiently arrange limited workers and snow removal machines. One of the methods to solve the problem is to measure the distribution of road snow depth in the area. By sharing the measured data, it can be used as a guide to advance dispatch, and efficient staffing and machine layout can be achieved.
In this study, we focused on an in-vehicle LiDAR equipped with multiple laser scanning planes as a sensor to measure the snow depth on the road, and carried out a mobile measurement experiment with the on-board LiDAR mounted on the vehicle to acquire the snow depth on the road. As a result, it was confirmed that the MAD of the snow depth measured by LiDAR and the snow depth measured by the snow scale was 3 cm, and the correlation coefficient between the three-dimensional point cloud generated by SfM and the cross-sectional shape was 0.98 or more. We also studied and implemented a method to generate a snow depth distribution from the measured 3D point cloud by an efficient method. As a result, high reproducibility was confirmed from the method comparing with the measurement data without snowfall. It is confirmed that the generation of the snow depth is important, although the road surface position is correctly determined. In the future, positioning by GNSS etc. will be performed simultaneously with LiDAR measurement, and the position information will be added to the three-dimensional point cloud to determine the road surface position. It was suggested that the snow depth could be generated with higher accuracy if the discrimination became easier.