Monitoring and Measurement of an Artificial Landslide Dam Using UAV Images and Image-Based Modeling

Abstract

Because of heavy rains or earthquakes, landslide and debris flows may flow into rivers and block river channel so to become a landslide dam. Due to the relatively loose nature and the absence of proper engineering spillway, landslide dams are often vulnerable to catastrophic failure by overtopping and breaching, and cause severe flooding in downstream area possibly with serious casualties, such as life loss and property damage. Therefore, the technology development of time-effective monitoring landslide dams becomes an important issue in disaster mitigation. However, the monitoring and assessment of landslide is often difficult due to the timeliness and effectiveness over an unreachable location. Since the timing of dam failure and the magnitude of the resulting flooding downstream are highly affected by the landslide dam size and geometry, the three-dimensional measurement of landslide dam is an essential task in the mitigation act following the occurrence of a landslide lake. Taking responsiveness and efficiency into consideration, this research proposed a process for the monitoring and assessment of landslide dam by employing unmanned aerial vehicles (UAVs) for aerial surveying and Image based modeling (IBM) for aerial surveying and a 3D model reconstruction of an artificial landslide dam built in a real Landou stream within Hui-Sun experimental forest in Nantou, Taiwan, as a full-scale on-site experiment. The result shows that a 3D terrain model can be successfully constructed based on the UAV images through IBM process. Especially, the volume and stability of landslide dam is the most urgent information needed to be acquired for mitigation strategy making so that the failure of landslide dams due to dam breach and erosion by the overflow stream can be predicted and prepared.