Abstract:

Coral reef structural complexity supports habitat and ecological function, so accurate measurement is essential for monitoring and restoration. In-water surveys face time, logistical, and safety constraints, creating a need for scalable approaches such as large-area imaging (LAI) and Structure-from-Motion (SfM) photogrammetry. NOAA needs consistent workflows to monitor restoration areas, integrate imagery from different platforms, and distinguish ecological change from methodological variability. Using the SfM section of NOAA’s Mission: Iconic Reefs LAI guidance as a reference, this NERTO project developed a reproducible framework for comparing diver- and ROV-based imagery while quantifying processing variability. The analysis used imagery from reef restoration sites, a relevant setting for fine-scale structural monitoring. We processed images in Agisoft Metashape using different alignment, depth-map quality, and filtering configurations. Reconstructions were screened using checkpoint RMSE, reprojection error, coverage, point density and confidence, completeness, noise, processing time, and file size. We compared leading products as standardized digital elevation models in R using a common region of interest and surface complexity ratio. Higher-detail reconstruction produced approximately four times more points and a larger file but changed surface complexity by only about 3%. Preliminary diver–ROV results showed broadly similar structural patterns, but the areas were not spatially paired and therefore cannot demonstrate platform agreement. These findings provide NOAA with a stronger QA/QC basis for selecting efficient processing settings and designing paired comparisons that support reproducible, scalable coral reef restoration monitoring.

The results presented are from the NOAA EPP CSC NERTO graduate internship conducted under the mentorship of Dr. Shay Viehman at NOAA’s NCCOS Beaufort Laboratory, NC. This project supports the NOAA CESSRST-II mission of understanding and predicting environmental changes, particularly in oceans and coastal ecosystems. The project investigated how variations in SfM processing parameters and diver- versus ROV-based imagery influence the accuracy, consistency, and efficiency of coral reef structural measurements, with the goal of developing standardized workflows for reproducible coral reef restoration monitoring.