Monitoring Excavation Slope Stability Using Drones

With its mobility and high visual data acquisition capability, drones can be used to capture videos for monitoring progress and safety for a variety of construction processes. By reconstructing three-dimensional (3D) point clouds from video images, drones can play an important role for quantitative analysis of productivity and safety-related metrics, especially in situations where construction worksites are rapidly evolving in response to construction operations. In this paper, we explore the usage of a drone to obtain videos of an excavation project in order to monitor slope stability at different stages of the operation. First, the drone is flown over the construction site to collect videos of the ongoing excavation. Then, the structure from motion analysis is performed on the video images to obtain dense 3D point clouds. The terrain points are separated from non-terrain points by detecting elevation changes and clustering smooth surfaces. Based on the terrain points, the surfaces of interest are extracted and their plane parameters are computed in order to estimate the slope stability of the evolving excavation. In addition, the digital elevation models (DEM) and the corresponding slope maps are generated from the terrain points to visually depict the progress of the excavation. Experimental results demonstrate the feasibility of the proposed method and show that a texture-rich 3D model of an excavation can be constructed from drone imagery and subsequently used for quantitatively evaluating slope stability and safety.