Automatic slope modelling accuracy assessment using point cloud data from unmanned aerial vehicle

Military commanders often determine appropriate routes for combat mobility from source to destination based on a prior understanding of the operating region's important topographic features and environmental factors. With advances in military technology, the types of maps needed for military operations have changed. The computer-based geographic information system (GIS) is an essential technology to replace current topographic surveys. Obtaining survey data with the current method is costly and creates hazardous working conditions, especially on steep slopes or remote areas. Unmanned Aerial Vehicle (UAV) technical capabilities performance has improved yearly. In photogrammetry, this advancement presents broader opportunities to obtain digital images to create large-scale maps. This technology can overcome hazards while reducing the time and cost of obtaining site data at the slope area. In this study, the Digital Surface Model (DSM) and UAV-derived path profiles of various heights are compared, and the accuracy of the data with and without using Ground Control Points (GCP) is examined. A total of seven GCPs and ten control points (CPs) were established in the study area. The study area is located within the National Defence University of Malaysia campus. DJI Phantom 4 captured aerial images from three different altitudes, namely 60m, 80m and 100m. All acquired images were processed using Pix4Dmapper software, and the products produced included digital orthophoto and DSM. Invalidating the accuracy of the produced DSM, the coordinates of the GCP of the survey line in the imaging area were extracted from the generated DSM using Global Mapper software. Compared to 0.43m for 100m without GCP, the average Root Mean Square (RMS) for 100m with GCP is 0.02m. Many inaccuracies are reduced by the GCP 100m altitudes. In conclusion, using GCP can deliver greater accuracy and attain the precision required for surveying. The 100m altitude point cloud data is suitable for mapping and modelling the slope or hilly area with reliable accuracy.