AI-Paired Infrared Cameras Contribute to Marine Mammal Conservation

The vessel-mounted Marine Observer scans the horizon, day or night, with a wide field for real-time marine mammal or object detection. (Image credit: Toyon)

Industrial operations can adversely affect marine mammals through exposure to loud acoustic sources and risk of vessel strikes. Marine mammals can be difficult for vessel operators to see, especially during the night or when high glare is present. On the West Coast of the United States, more than 80 large whales are killed annually by vessel collisions. Marine Observer, designed by Toyon Research Corporation, is a tool developed to mitigate the risks to marine mammals. This new technology has been used by the National Oceanic and Atmospheric Administration (NOAA), and by the oil and gas and offshore wind industries to support scientific census efforts, vessel strike avoidance, and mitigation of acoustic impacts on whales.

Novel Technology

Marine Observer’s cus¬tom-built camera houses multiple long-wave infrared (LWIR) camera cores and an inertial navigation system for an electronically-stabilized and seamless field of view covering up to 90° per camera. The supporting software processes the video stream in real time and uses artificial intelligence (AI) to automatically detect whale blows and vessels. Operators are alerted to the presence of whales, and sightings are automatically archived. A human operator makes the final decision regarding the pres¬ence or absence of marine mammals. An additional benefit of Marine Observer is that it can provide a live, stabilized video stream that can be used to watch for objects at night that might pose a collision risk to the vessel and/or entanglement risk to towed gear. The camera is easy to use, the software is intuitive, and it can be operated by a single analyst.

Whale Blow Marine Observer

A screenshot from the Marine Observer detection software. The red box highlights an automatic detection of a blow (spout) from a gray whale seen in the image. (Image credit: Toyon)

Leveraging AI

Humans are easily fatigued when staring at a display or out at the ocean for extended periods of time. By automating the detection process and cueing an operator to the location of an unconfirmed whale sighting, it streamlines the human’s role to the relatively simple process of confirming or denying its presence. This allows the system to perform what machines do well—carry out tedious, repetitive tasks—and allows the humans to do what they do well—make exquisite classification decisions. Toyon collects more data with each deployment, allowing the team to continue to improve the AI algorithms and further reduce the workload of the operator.

To find out more about Marine Observer, visit: www.marineobserver.com

This spotlight appeared in Environment, Coastal & Offshore (ECO) Magazine’s 2023 Winter edition Fisheries & Aquaculture, to read more access the magazine here.

latest issue

Sampling and survey techniques, methods, and technologies are far-reaching across the applied marine science space. These days, many instruments and…

Search