Change Detection of Landslide-Prone Terrain on Steep Mountain Roads Using Drone and Satellite Imagery

In this study, a quantitative change detection analysis of surface sliding phenomena was conducted in a landslide-prone area near the Piam Tunnel in Dogye-eup, Samcheok-si, Gangwon Province. Multi-temporal UAV imagery was integrated with satellite data from 2015, 2017, and 2022. ArcGIS-based raster differencing showed that approximately 95% of changes fell within ±2 standard deviations, indicating overall stability across most of the area, whereas the remaining 5% of extreme values exhibited strong spatial correspondence with actual landslide hazard zones. The time-series comparison revealed that the 2022–2015 period recorded the highest standard deviation (57.44), reflecting the most pronounced geomorphic changes, while the 2017–2015 period showed comparatively lower variability (48.02). These findings highlight that structural countermeasures alone are insufficient for slope hazard mitigation. Continuous drone-based monitoring combined with multi-temporal change detection provides a robust and practical framework for quantitative landslide risk assessment and proactive hazard management.