Improving Data Acquisition Efficiency: Systematic Accuracy Evaluation of GNSS-Assisted Aerial Triangulation in UAS Operations

Abstract

The number of unmanned aerial system (UAS) mapping applications that require high-accuracy data sets have grown significantly in recent years. Indirect georeferencing (IG) requires the establishment of multiple, well distributed ground control points (GCPs) throughout a project site. UAS mapping operations can now use precise navigation trajectories with position and/or orientation observations to replace or augment GCPs. Assisted aerial triangulation (AAT), which uses only airborne camera perspective center (CPC) observations to augment GCPs, is the primary method investigated herein. The objective of this study is to rigorously analyze the impact that each additional GCP in UAS structure from motion (SfM) photogrammetric processing has on the spatial accuracy of UAS-derived orthophotos and digital surface models (DSMs) when comparing AAT to IG. Three separate flights with two different cameras were flown over a project site that had a network of nine GCPs and 70 additional checkpoints (CPs). The authors carried out 29 unique processing trials for each flight, varying the georeferencing method and the quantity of GCPs used. American Society of Photogrammetry and Remote Sensing’s (ASPRS’s) Positional Accuracy Standards for Digital Geospatial Data provided reporting metrics to evaluate the derived orthophoto and DSM from each processing trial. UAS SfM processing using AAT with minimal to no augmenting GCPs provided high spatial accuracy products that met or exceeded more traditional workflows reliant on IG. When using AAT, the inclusion of at least one GCP improved vertical spatial accuracy relative to processing trials by using airborne positioning alone. To ensure the focal length was resolved during camera self-calibration, the inclusion of at least two high-accuracy GCPs in the simultaneous bundle block adjustment (SBBA) is recommended.