Perry Naughton, Clinton Edwards, Vid Petrovic, Ryan Kastner, Falko Kuester and Stuart Sandin. 2015.

Read the full article here.

ABSTRACT:

Visually documenting seafloor habitats has the potential to answer challenging questions in several maritime disciplines including: ecology, geology, and archaeology. Unfortunately, the attenuation of visible light underwater limits the imaging footprint of a single image to square meters. This limitation makes representing large habitats, on the order of hundreds of square meters and beyond, an intensive process requiring the collection, storage, processing and annotation of thousands of high resolution images per hundred square meters of seafloor. This paper describes a pipeline for dealing with these challenges efficiently and effectively using visual data of coral reef communities processed into a three dimensional model. We evaluate the resources and technological advancements required to scale this problem to orders of magnitude larger than the current state of the art and motivate the need for networked underwater data collection platforms to push the scalability of this method.