Remote Sensing Approach to Riprap Slope Inspection

Abstract Only - Riprap, or rock fragments, armor many of the upstream and downstream slopes of dams, dikes, and levees. The riprap must withstand the forces of wind, rain, and waves. Thus, the material must be adequately heavy and durable to protect the slope from these erosive forces. Over time, the riprap can degrade by breaking into smaller pieces. Therefore, periodic inspection is important to confirm that the riprap remains sufficiently heavy to resist wave action. Current procedures for inspecting riprap involve a trained dam inspector subjectively evaluating present conditions and comparing their observations to notes and photos from previous inspections. Shortcomings of the current inspection practices include the time-intensive nature of the work and the reliance on subjective evaluations, which may be inconsistent among different inspectors. The present study presents an empirical relationship between the surface roughness of three-dimensional (3D) point cloud data captured of the riprap slopes and the median size, D50 of the riprap material, which allows for an objective assessment of rock size. The point cloud data, a dense collection of 3D position measurements, can be generated using lidar or photogrammetry. Lidar is an active remote sensing technology that maps the 3D environment by emitting laser pulses. Structure from motion photogrammetry uses sophisticated computer software to refine the structure of the 3D environment using a set of photographs collected by handheld camera or autonomous drone. Field studies are an important component of this project; however, physical sampling of the riprap at the surface of a slope only discloses an estimate of the true particle size distribution. Therefore, a simplified kinematic model of 2D disks was used to test the empirical relation between roughness and rock properties. This model randomly generates beta-distributed assemblages of disks with the following prescribed parameters: maximum, minimum, median disk diameter by disk area, and coefficient of uniformity, Cu. In addition to numerical simulations, empirical correlations have been developed for three separate case studies. For each case study, method success and lessons learned are presented.