NSF Award
2316801
Coral reefs are one of the most diverse ecosystems on Earth, hosting at least 30% of all marine species. They generate almost $10 trillion USD per year globally and provide protein and income for more than 5% of humanity. Unfortunately, most coral reefs are in severe decline due to the climate change and local stressors. Some reefs, however, appear to be more resilient than others. We need to find and identify these reefs quickly to elucidate the factors that confer this resilience. Historically, the kinds of surveys required to locate and demonstrate resilience have been slow, expensive, and reliant on highly trained scientists. These traditional methods are too slow and too cumbersome to provide 21st Century solutions for our 21st Century challenges. In this project, advanced 3D computer vision and photogrammetric techniques will be assembled into a novel computational toolset called CoralReef3D, which will be made available to divers and coastal surveyors all over the world. This shareware will allow non-experts to create georeferenced (and therefore repeatable) 3D reef montages that can easily be assessed for benchmark studies on coral growth and survival. The impact of this research will be broadened by open-source software, online workshops, student training, and outreach activities involving museums and the general public.<br/><br/>CoralReef3D is expected to bring forth significant advancements in underwater 3D measurements. During this research, novel water-refraction correction methods will be developed and incorporated into CoralReef3D. Correction for water-refraction distortions at the camera/water interface will allow production of accurate measurements of seascape montages. For instance, the program’s advanced capabilities for object-recognition and semantic-segmentation will allow it to generate on-image scales for linear or volumetric measurements. This will also enable precise calculations of ecosystem rugosity (topological complexity) and geomorphology in surveyed areas. By providing GPS information on all images, CoralReef3D can fate-track individual colonies and quantify ecosystem changes over time. Both controlled laboratory validation and in situ field validation will be conducted. Controlled laboratory validations will be performed in deep seawater tanks maintained by Reef Systems Coral Farm. In situ field validations will be performed in Hawaii and the Florida Keys along depth gradients. To shorten the learning curve normally associated with sophisticated image analysis programs, CoralReef3D will be extensively documented using both text and pictorial instructions. These technological advances will produce a benchmark method for 3D modeling of complex objects in water. CoralReef3D will empower both experts and non-specialists to make affordable, large-scale, and accurate assessments of rapidly-changing marine ecosystems.<br/><br/>This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
