Skip to main content

Initiative #49

← AI initiative registry

Salmon Fence Counting​

Listed in the registry Pilot Original language : English

About the initiative

Description of the initiative

A model using computer vision to analyze water stream fence video-footage, to simultanously count the number of and predict the species of salmon.​ YOLO 11 Chinook trained on 735 annotated frames from 63 videos Coho trained on 432 annotated frames from 44 videos Sockeye trained on 1149 annotated frames from 36 videos Chum untrained (annotation of 1368 frames from 43 videos in progress) Pink Humpback

AI family

Perception and Understanding

Bucket rationale

Perception and Understanding

Lifecycle

  1. Ideation

    Stage status : Completed

    May 29, 2025

  2. Assessment

    Stage status : Completed

  3. Approved

    Stage status : Completed

  4. Development

    Stage status : Completed

  5. Pilot

    Stage status : In progress

  6. Production

    Stage status : Upcoming

  7. Archived

    Stage status : Upcoming

Schedule for this initiative

  • Completed
  • In progress
  • Planned
  • Not applicable
  • Today
2025 May Jun Jul Aug Sep Oct Nov Dec 2026 Jan Feb Mar Apr May Jun Jul Aug Sep
  1. Ideation

A similar need?

Nobody has come forward yet. If your team faces the same problem, say so: it helps bring initiatives together and share a solution.

Your email is never published: only the AI Enablement team sees it, to connect you.

Contacts

Requester
Quan, Eric
Sector contact
Quan, Eric
Subject matter expert
Quan, Eric

Data and tools

Sector
Programs Sector
Region
Pacific
Could this initiative be shared with the Treasury Board Secretariat?
Yes
Submitted on
May 29, 2025

Primary users

  • Departmental employees

Government priorities and declaration

Declared to TBS Recorded as transmitted to TBS; not editable here.

Declared name of the AI system
PSSIFence
TBS registry identifier
2526-DFO-MPO-007
Status declared to TBS
Experimentation
Sent to TBS on
October 10, 2025
Purpose of the system, as declared
Problem: DFO inspects hours of underwater camera footage to count and identify fish species through fence counts, which is a structure installed across a river or stream to monitor fish population. The amount of hours needed to review underwater footages will need to increase in order fulfil oblgiations as it is not sustainable. Objective for AI: Replace manual analysis of fence videos to determine salmon counts by species. Leveraging advanced computer vision technology, reducing the manual effort required, and providing more accurate and timely data.