The eastern oyster is an economically important North Carolina fishery, but due to monitoring challenges, a traditional stock assessment has never been conducted. Unmanned aircrafts aid in overcoming hurdles and are increasingly implemented to map estuaries; they provide higher resolution than satellites with less intensive ground-truthing methods. Drone mapping has unique challenges including the image distortion, tide levels, and placing accessible ground control points (GCP). This project is designed to better understand the new technology’s capabilities, accuracies and limitations. The project assesses the horizontal geospatial error as a result of wind conditions, sunlight levels, flight altitude, and tide levels and how these factors affect the oyster reef habitat and abundance estimates. Oyster reef in Morehead City was imaged, using the surrounding docks as permanent GCPs. A DJI Phantom 4 Advanced, including a GPS/GLONASS INS with ±1.5 m horizontal accuracy positioning and a camera with a GSD (ground sampling distance) of 1.67 at 60 meter flight altitude, was used for the project’s duration. An orthomosaic was generated, first with direct georeferencing and then with naturally-occurring GCPs. Paired with relative abundance predictions, aerial maps can determine current reef footprint and estimate oyster abundance, which aids fisheries management.
Results/Conclusions
Following the ASPRS Geospatial Data Standards, a RMSE (root mean square error) was calculated by comparing map-derived check point coordinates to the GPS coordinates collected using a Real-Time Kinematic (RTK) device. This method was used to quantify horizontal accuracy of each map. Preliminary calculations suggest that flying at higher altitudes not only decreases the ground resolution, but increases the RMSE. Weather conditions, such as cloud coverage and wind, do affect RMSE calculations and should be considered when determining the optimal time to fly. While tide level had less of an effect on RMSE values, the image distortions due to high water levels might hinder photo alignment and other analysis techniques such as object-based image analysis (OBIA). Generally, the RMSE values calculated for the maps generated with no GCPs were higher than ASPRS Standards. The average resulting error may be acceptable for reef footprint estimates and further analysis will be conducted to determine how this alters oyster relative abundance predictions. It is important to assess the limitations and accuracy of drone aerial mapping as unmanned aircraft generated maps will be increasingly used to inform stock assessments and fisheries policy.