An Approach to Filter Elevation Errors in UAV LiDAR Scanning on Open Pit Mine Slopes

As a remote sensing tool, LiDAR has advantages of high resolution, high accuracy, and the ability to mitigate the effects of visual obstructions, like vegetation, on measurements. LiDAR, when mounted on unmanned aerial vehicle (UAV), is now widely employed in geodesy, mapping, resource exploration, and risk assessment. In open-pit mines, UAV LiDAR supports regular scanning of large, hard-to-reach areas and continuous data collection to track slope movement over time. While it has been utilized for monitoring slope stability, its effectiveness in measuring slope deformation in open-pit mines is significantly impacted by the steep slope gradient, which can lead to increased vertical detection errors and create noise points in the generated point cloud. This paper introduces an approach to identify the elevation errors based on ground references and a new point cloud data filtering algorithm to remove these noise points. This method can improve the accuracy of UAV-based LiDAR applications in steep slope environments, and will help it become a reliable tool for early warnings of potential instability in steep bench slopes and overall inclined slopes.