1:45 PM - 3:15 PM
[HDS08-P09] Monitoring of large-scale landslide activity using repetitive RTK-UAV photogrammetry -a case study in Nigorizawa landslide on the south-eastern Oze-
Keywords:UAV-SfM, GNSS, erial photograph interpretation, differential analysis, micro-landforms
In order to capture small-scale mass movement phenomena and their precursor microtopographic changes in the interior of large-scale landslides, and to evaluate landslides over time and over an area for the purpose of contributing to mountain land use, especially maintenance and management of structures in the interior of landslides, photogrammetry using RTK-UAV has been continuously conducted on large-scale landslides with power transmission towers installed in the interior. The research target is a large landslide in the southern part of Ozenuma swamp. The research target is a large-scale landslide with a width of about 1 km, located about 3 km southeast of Ozenuma. On the east side of the landslide, there is a power transmission tower line running in a north-northeast to south-southwest direction. Aerial photo interpretation of several periods indicates that landslide movement has continued from 1950's to the present. To prevent destabilization due to landslide movement, the foundation of the steel tower located in the southeast part of the landslide has been reinforced. The area around the pylon road has been cleared for a width of approximately 100 m, which facilitates UAV photogrammetric surveying of the landslide interior.
UAV surveying was conducted four times since November 2020, mainly in October-November just before snow accumulation and in June just after snow melt. Photogrammetry was conducted using DJI's Phantom4RTK and D-RTK2 mobile stations, and 3D modeling was conducted using Agisoft Metashape Profesional. 10 cm DSM and 3 cm high-resolution orthoimages were acquired at each time period. To verify the accuracy of the 3D model, several validation points were set up in the logging area. The coordinates were obtained by PPK analysis using u-blox ANN-MB-01 and Qwiic GPS-RTK2 module with ZED-F9P and the electronic reference point "Katahina". The difference between the photogrammetric and GNSS positioning at the verification points was ±0.1 to 0.4 m in the horizontal direction and less than +1.4 m in the vertical direction. The DSM difference analysis indicates that a terminal collapse of about 4 m deep and 10 m wide occurred in a small landslide block in the center of the logged area during the snowmelt season in 2022. In addition, we confirmed that the landslide fluctuation has been continuing an annual basis, such as the area of 0.2 m/year of areal subsidence in a sparsely vegetated area with a continuous distribution of horseshoe-shaped depressions. We will continue to monitor microtopographic changes and investigate the risk of damage to the tower road by continuing UAV photogrammetric surveying and GNSS surveys of the entire large-scale landslide.
UAV surveying was conducted four times since November 2020, mainly in October-November just before snow accumulation and in June just after snow melt. Photogrammetry was conducted using DJI's Phantom4RTK and D-RTK2 mobile stations, and 3D modeling was conducted using Agisoft Metashape Profesional. 10 cm DSM and 3 cm high-resolution orthoimages were acquired at each time period. To verify the accuracy of the 3D model, several validation points were set up in the logging area. The coordinates were obtained by PPK analysis using u-blox ANN-MB-01 and Qwiic GPS-RTK2 module with ZED-F9P and the electronic reference point "Katahina". The difference between the photogrammetric and GNSS positioning at the verification points was ±0.1 to 0.4 m in the horizontal direction and less than +1.4 m in the vertical direction. The DSM difference analysis indicates that a terminal collapse of about 4 m deep and 10 m wide occurred in a small landslide block in the center of the logged area during the snowmelt season in 2022. In addition, we confirmed that the landslide fluctuation has been continuing an annual basis, such as the area of 0.2 m/year of areal subsidence in a sparsely vegetated area with a continuous distribution of horseshoe-shaped depressions. We will continue to monitor microtopographic changes and investigate the risk of damage to the tower road by continuing UAV photogrammetric surveying and GNSS surveys of the entire large-scale landslide.