Accuracy Assessment of Terrestrial and Drone LiDAR-Based Model for Disaster Site Investigation

This study aims to evaluate the applicability of ground and drone light detection and ranging (LiDAR) for use in disaster sites where damage to facilities and structures has occurred.For this purpose, 3D data was acquired from a disaster environment building of the Korea Institute of Robot and Technology Convergence, using Terrestrial LiDAR for indoor areas and drone LiDAR for outdoor areas.The indoor data were aligned using different methods (target registration, auto, and visual alignment), while the outdoor data were modeled at two different altitudes (30 m and 50 m), resulting in two separate 3D models for each scenario.An accuracy assessment of each model was conducted, revealing that the indoor model's root mean square error (RMSE) was approximately 2 cm, while the outdoor model's RMSE was around 3 cm.Additionally, indoor and outdoor models were aligned across eight different cases, with accuracy evaluations carried out accordingly.The results showed no significant differences based on the alignment method.The best accuracy was observed in the alignment model between the target registered indoor model and the 30 m altitude outdoor model, with RMSE values of 0.020, 0.023, and 0.018, respectively.In contrast, the auto and visual alignment indoor model combined with the outdoor model at a 30 m altitude showed the lowest accuracy.Based on these findings, it is anticipated that diverse investigations can be conducted by integrating ground and drone LiDAR data, depending on the type of disaster and site conditions.