Unmanned aerial vehicle data acquisition methodology and assessment of photogrammetric products for telecommunication mast inventory

Abstract UAV (unmanned aerial vehicle) data are widely utilized for the inspection of telecommunication masts, but the quality and accuracy of photogrammetric products largely depend on the method of data acquisition and flight parameters. This study proposes a methodology for flight trajectory planning and acquiring UAV imagery for inventorying telecommunications masts. A crucial component of the methodology is an application that facilitates flight mission planning and enables operators to collect reliable and repeatable data fast. An accuracy analysis was performed for both dense image matching (DIM) point clouds and 3D mesh models to evaluate quality of products with terrestrial laser scanning (TLS) point clouds used as the reference. A similar evaluation was performed with unmanned laser scanning (ULS) data. The experiments demonstrated that the mean distance between the DIM point cloud and the TLS data was not higher than 2 cm and standard deviation was 1.8 cm or less for all surveyed towers while for 3D mesh models these values were less than 1.2 cm and 2.6 cm respectively. A comparative analysis with ULS data also confirmed that both the DIM point cloud and the mesh model meet the accuracy criteria adopted in the research project.