In the effort to preserve salmon populations in rivers across British Columbia, the use of artificial intelligence (AI) proves to be a valuable tool. Instead of relying on traditional methods like volunteer teams or river footage, AI technology has been […]
In the effort to preserve salmon populations in rivers across British Columbia, the use of artificial intelligence (AI) proves to be a valuable tool. Instead of relying on traditional methods like volunteer teams or river footage, AI technology has been developed to utilize deep learning computer vision to identify species. A recent pilot study published in the journal “Frontiers in Marine Science” showcases the “Salmon Vision” computer vision model, which achieved an accuracy rate of around 90% in recognizing coho salmon and 80% in identifying sockeye salmon by analyzing over 500,000 video clips.
A team of researchers has developed and trained a computer vision model that utilizes AI to examine video images of salmon, accurately identifying species and providing precise fish counts. This AI technology replaces the role of technicians who would traditionally review video footage for four to five months.
The data for this study was collected from weir structures built in the Skeena River watershed. A weir is a traditional device constructed by Indigenous peoples for capturing fish at the river’s edge and was used for monitoring during the pilot project. The Heiltsuk First Nation and the Wild Salmon Center joined forces to pioneer the use of AI for real-time salmon monitoring in British Columbia.
This AI technology enables the collection of high-quality data that will help the Heiltsuk Nation make better decisions and have stronger management, particularly regarding the management of sockeye salmon on their territory. Considering the decline in salmon numbers over the past few decades and the lack of data, AI allows for faster data analysis and provides critical information that can aid these vulnerable fish. The project is funded through the British Columbia Salmon Restoration and Innovation Program, implemented by the Canadian Ministry of Fisheries and Oceans.
Frequently Asked Questions:
- What is the role of artificial intelligence in salmon counting?
- How are salmon traditionally counted?
- Where were the data collected for the pilot study?
- What are the advantages of using artificial intelligence in salmon counting?
- How will artificial intelligence help the Heiltsuk Nation?
Artificial intelligence is used for species recognition and accurate counting of salmon based on video footage.
Salmon are traditionally counted by volunteer teams or through river footage.
The data for the pilot study was collected from weir structures built in the Skeena River watershed.
The advantages of using artificial intelligence include faster data analysis, more precise species recognition, and the ability to monitor salmon numbers in real-time.
Artificial intelligence will assist the Heiltsuk Nation in collecting quality data and making better decisions regarding the management of salmon populations on their territory.