3D Classification of Cold-Water Coral Reefs: A Comparison of Classification Techniques for 3D Reconstructions of Cold-Water Coral Reefs and Seabed

Submarine canyons and cold-water corals (CWC) are key components of ocean ecosystems due to their complex geomorphology and their role in nutrient supply dynamics. As three-dimensional structural habitats, there is a need for robust and accessible technologies to enable more accurate representations and assessments of these environments. The use of Remotely Operated Vehicles (ROV) and Structure-from-Motion (SfM) photogrammetry combined represent an effective and non-destructive method that yields high-resolution reconstructions of deep-sea environments such as CWC habitats and submarine canyons. This research focused on the application of underwater photogrammetry to improve seabed classification techniques on the Porcupine Bank Canyon (PBC), located at approx. 300 km offshore Ireland. We have developed a classification scheme in different habitats within the canyon which will aid the understanding of the main environmental patterns of CWC distribution within the canyon. Three binary classification methods were applied, namely: Multiscale geometrical classification (MGC), Colour and Geometrical classification (CGC) and Object-based classification (OBIA) for the baseline method. The application of these methodologies brought novelty to the study because, although these approaches are well stablished in other fields such as terrestrial image analysis, they had not yet been applied to marine environments. We analysed the classification accuracy results, overall performance and the potential loss of information when using 2D and 3D data. Our results showed that 3D the MGC method outperforms the CGC methods. However, each method has advantages for specific applications pertinent to the wider marine scientific community. Results suggest that advancing from commonly employed 2D image analysis techniques to 3D photogrammetry classification methods is advantageous and may provide a more realistic representation of CWC habitats. Key-words: Cold-water corals, submarine canyons, 3D photogrammetry, 3D image classificatio