Isana Imai

A Study on Estimation of Light Environment in Forest Using Drone LiDAR Point Cloud

Kazuyoshi Takahashi

According to the Forestry Agency, forests have multifaceted functions, and forest maintenance is conducted to maintain these functions. In order to determine the degree of effectiveness of forest improvement, monitoring surveys are conducted. Part of the monitoring survey is to investigate the light environment in the forest, and it is thought that a drone LiDAR point cloud can measure a wide area efficiently and safely. Therefore, we investigated the possibility of using drone LiDAR point clouds to estimate the light environment in forests.
We measured the open sky ratio using a 360-degree camera over a five-sided waterside area in Ojiya City, Niigata Prefecture. We measured a drone LiDAR point cloud in the same area, and calculated an openness index based on the same approach as that used to calculate the openness index from the LiDAR point cloud. Based on these two sets of data, the measured values of the air openness index and the calculated values of the air openness index are calculated. If the difference between the measured sky coverage and the calculated sky coverage index is small, it is considered possible to estimate the light environment in the forest using the drone LiDAR point cloud.
The correlation between the openness index and the openness rate was investigated by changing the calculation method of the openness index. We considered that the laser beam reaches the ground when the LiDAR laser beam is within the 60-degree zenith angle range, so we used data within the 60-degree zenith angle range for both the calculation of the air openness rate and the air openness index. The results showed that the pitch angle of the drone did not have a significant effect, and the magnitude of the error varied depending on the voxel size, which was determined when calculating the air coverage index. In this study, the voxel size was varied from 0.01m, 0.2m, and 0.5m to calculate the air openness index, and it was found that the correlation was highest when the voxel size was 0.5m.
In conclusion, it is possible to estimate the light environment in forests by using drone LiDAR point clouds. However, it is necessary to consider the appropriate voxel size and to study different vegetation environments in the future.