by A team of scientists, natural resource managers, and conservation organizations from Canadian First Nations, governments, and academic institutions have published the first results of a groundbreaking salmon population monitoring tool called Salmon Vision. Combining artificial intelligence with traditional fishing weir technology, Salmon Vision has demonstrated remarkable accuracy in identifying and counting fish species, potentially revolutionizing real-time salmon population monitoring for fisheries managers.
Lead author, Dr. Will Atlas, Senior Watershed Scientist at the Wild Salmon Center based in Portland, states that in recent years, underwater video technology has shown promise in visually tracking the return of salmon to rivers. This aligns with the feedback received from First Nations partners, who emphasize the need for automated fish counting to make informed decisions while salmon are still migrating.
The Salmon Vision pilot study focused on two Indigenous-run fish counting weirs located on the Kitwanga and Bear Rivers of British Columbia’s Central Coast. The project annotated and analyzed over 500,000 individual video frames using a deep learning computer model developed in collaboration with the Gitanyow Fisheries Authority and Skeena Fisheries Commission. The model exhibited promising accuracy, achieving mean average precision rates of 67.6% in identifying 12 different fish species passing through the fish-counting boxes at the weirs. Notably, the scores for coho and sockeye salmon, two principal fish species targeted by First Nations, commercial, and recreational fishers, surpassed 90% and 80%, respectively.
Dr. Keolu Fox, a professor at the University of California-San Diego, expressed enthusiasm for the project, highlighting its alignment with the goals of supporting Indigenous futurism and climate resilience. Dr. Fox was among several early reviewers who contributed to the crowdfunding efforts for the development of Salmon Vision.
Real-time data on salmon returns is crucial for multiple reasons. According to Dr. Atlas, many fisheries in British Columbia have lacked adequate data for decades, resulting in fisheries managers relying on early-season catch data rather than the actual number of salmon returning. In addition, changing weather patterns, stream flows, and ocean conditions have led to more variable salmon returns, exacerbating the risks of overfishing already vulnerable populations. By collecting and analyzing data in real-time, Salmon Vision addresses the information gap and facilitates the development of climate-smart and responsive fisheries.
Dr. Atlas emphasizes that Salmon Vision will be an invaluable tool for First Nation fisheries managers and other organizations. It will not only provide better information for managing conservation risks and fishing opportunities at the decision-making table but also assist in remote rivers across salmon country, where on-site data collection is challenging and costly.
The Salmon Vision team plans to implement automated fish counting on a trial basis in several rivers along the B.C. North and Central Coasts in 2023, ultimately aiming to provide reliable real-time count data by 2024. This innovative technology holds great promise for the recovery of salmon populations, paving the way for more informed and sustainable fisheries management.
*Note:
1. Source: Coherent Market Insights, Public sources, Desk research
2. We have leveraged AI tools to mine information and compile it
Ravina Pandya, a content writer, has a strong foothold in the market research industry. She specializes in writing well-researched articles from different industries, including food and beverages, information and technology, healthcare, chemicals and materials, etc. With an MBA in E-commerce, she has expertise in SEO-optimized content that resonates with industry professionals.
