Toward low-cost automated monitoring of life below water with deep learning

Abstract

Oceans will play a crucial role in our efforts to combat the growing climate emergency. Researchers have proposed several strategies to harness greener energy through oceans and use oceans as carbon sinks. However, the risks these strategies might pose to the ocean and marine ecosystem are not well understood. It is imperative that we quickly develop a range of tools to monitor ocean processes and marine ecosystems alongside the technology to deploy these solutions on a large scale into the oceans. Large arrays of inexpensive cameras placed deep underwater coupled with machine learning pipelines to automatically detect, classify, count, and estimate fish populations have the potential to continuously monitor marine ecosystems and help study the impacts of these solutions on the ocean. In this paper, we successfully demonstrate the application of YOLOv4 and YOLOv7 deep learning models to classify and detect six species of fish in a dark artificially lit underwater video dataset captured 500 m below the surface, with a mAP of 76.01% and 85.0%, respectively. We show that 2,000 images per species, for each of the six species of fish is sufficient to train a machine-learning species classification model for this low-light environment. This research is a first step toward systems to autonomously monitor fish deep underwater while causing as little disruption as possible. As such, we discuss the advances that will be needed to apply such systems on a large scale and propose several avenues of research toward this goal.

Monitoring data for a new large offshore marine protected area reveals infeasible management objectives

“…Moreover, imagery surveys are less invasive and can be conducted off a variety of platforms, making it more feasible to increase sampling effort and statistical power than trawl surveys. Current developments in seafloor imagery equipment (Clayton & Dennison, 2017;Dominguez-Carri o et al, 2021) and annotation (e.g., machine learning methods, Ayyagari et al, 2023;Piechaud & Howell, 2022), are expected to facilitate data collection and analysis. We acknowledge that our empirical data for sea pens are based on a single survey and do not capture temporal variation.…”