Advances in Information Technology Improve Open Science Use of Underwater Imagery

Underwater imagery (including video and still frame photos) are routinely collected with crewed and uncrewed submersible platforms and represent an unparalleled visual repository of knowledge about marine organisms and how they interact with their environment in a given spatiotemporal setting. Underwater images are of key importance for scientific analysis to inform policy and resource management decisions and are also in high demand for media and documentary releases. Underwater imagery requires accompanying metadata describing the contents of the imagery to achieve FAIR (Findable, Accessible, Interoperable, and Reusable) guidelines. This essential metadata generation step is an intense - often manual - process requiring scientists to provide video annotations that are a prerequisite to achieving Open Science initiatives and ensuring that data are available for broad exploitation in standard or common formats. Unique data management challenges set underwater imagery apart from most other oceanographic and geophysical data collections. Consistently high data volumes coupled with rapidly evolving technical advances in sensors can exceed data processing, management, and user capabilities. These challenges often require unique storage and data access solutions. These concerns are compounded by imagery collected from previous sea-going expeditions. These historical data - often decades old - take on a new importance in the face of climate change, and represent an opportunity to evaluate - visually and quantitatively - change over time for a specific organism or location. It is necessary that these historically significant records are retained and equally available for comparative analysis. Through the past two decades, the National Oceanic and Atmospheric Administration has researched, implemented, and refined processes that support best data management practices at minimal cost. NOAA Ocean Exploration has sponsored hundreds of ocean exploration missions since its inception in 2001. Ongoing collaboration between NOAA Ocean Exploration, the NOAA Central Library and the NOAA National Centers for Environmental Information has progressively defined best data management practices and developed tools to enhance user access to oceanographic data with a specific focus on underwater imagery. This collaboration resulted in a defined video metadata standard and a Video Data Management System to make analog video media available as a Library resource. Subsequently the Video Data Management Modernization Initiative modeled digital video access, based on entertainment industry standards which stored high resolution videos in deep storage with more easily accessible low-resolution proxies on spinning disk. Over the years, this model has been refined to manage imagery data of varying complexity as well as to leverage technological advancements to further improve data Fairness. We will offer an overview of several ongoing projects that leverage information technology advancements in cloud, artificial intelligence, machine and deep learning applications to benefit the imagery data analysis and management community by creating tools and algorithms in a crowd-sourced, open data manner. The capabilities of NOAA's imagery data management system have, and continues to, grown extensively thanks to broad collaboration with the larger ocean imagery data management and analysis community.