Q&A | HydroAI
HydroAI is Innovasea’s AI-powered fish monitoring system designed to improve the quality, quantity and transparency of fish passage data at hydropower facilities. The system combines high-resolution video cameras with proprietary, cloud-based software to deliver continuous, real-time information on fish movement in and around hydropower plants. By automating monitoring tasks that have traditionally required extensive human effort, HydroAI reduces the time and labour associated with manual fish counts and video review, while providing a consistent dataset that can be shared with regulators and stakeholders. The system is built around several integrated components. High-resolution cameras capture 24/7 footage of migrating fish, while a custom camera insert allows the hardware to be installed within a range of existing fish passage structures. An onsite edge box processes camera data and transmits it to Innovasea’s cloud-based platform, where proprietary algorithms analyse the footage and deliver results to a reporting dashboard. This architecture allows operators to access fish passage data remotely and in near real time. HydroAI is designed to integrate with existing infrastructure and support monitoring both upstream and downstream of dams. Innovasea positions the system for use at hydropower facilities undergoing licensing or relicensing, at sites operating under frequent regulatory scrutiny, and at locations where protected species or habitats may be affected. By improving data availability and transparency, HydroAI is intended to support environmental mitigation efforts while enabling more informed operational oversight at hydropower facilities.
www.innovasea.com
of error in the range of 30–40%, largely due to limited sampling and the need for interpolation.
Deployment at existing hydropower Below:HydroAI hub
facilities A key objective of the Species Aware project is ease of deployment. The system has been designed to integrate with existing fish passage infrastructure rather than requiring new construction. The camera insert is designed to fit into fish ladders with minimal modification. While some site-specific customisation may be required to accommodate different ladder dimensions, installation is straightforward. “We’re talking about a one- to two-hour installation,” Quirion said. “It does not affect the operation of a hydropower plant.” This approach minimises downtime and allows the
technology to be retrofitted at facilities of different ages and designs. The standardised camera positioning also supports consistent data collection across sites.
Operational benefits beyond
compliance Although regulatory compliance is a primary driver for fish monitoring, Innovasea identifies additional operational benefits associated with automated species identification. Some hydropower operators are already using
HydroAI to support day-to-day management of fish passage infrastructure. The system provides real-time video access to fish ladders, allowing operators to confirm flow conditions, identify obstructions and verify that passageways are functioning correctly. This capability is useful even outside peak migration periods, providing ongoing visibility into infrastructure performance. There are also potential reputational benefits. According to Quirion, operators using advanced monitoring systems are increasingly seen as demonstrating a proactive commitment to environmental
14 | January 2026 |
www.waterpowermagazine.com
responsibility. Looking ahead, automated monitoring could support more dynamic operational decision- making. Some hydropower plants face restrictions on generation during certain periods to protect fish, even when fish may not be present. “If you could know very precisely, on a minute-by-
minute basis, how much fish is passing by, perhaps this could drive operational decision-making,” Quirion said. He emphasised, however, that such applications depend on the technology becoming fully proven and trusted by regulators.
Improving ecosystem protection Reliable species-level data enables more targeted
approaches to ecosystem protection. By understanding which species are present and when they are moving, operators can take actions that are better aligned with actual environmental conditions.
“If you can have that knowledge with a lot of certainty, then you can take actions that are more effective,” Quirion said. This supports what he described as the ongoing challenge of balancing sustainable hydropower generation with protection of fish and their ecosystems.
Testing across diverse hydropower sites
As part of the Species Aware project, Innovasea will test the system at multiple hydropower sites with varying environmental conditions. These include differences in lighting, turbulence, debris loads and flow regimes. The aim is to evaluate both reliability and generalisation. Generalisation refers to the model’s ability to perform consistently across different sites rather than being optimised for a single location. The system will also be assessed for its ability to distinguish fish from debris, a common challenge in fish passage environments. Although initial training focuses on species found in North America, Innovasea has designed the system to be adaptable to international markets. Rather than training separate models from scratch, the company is developing a generic base model trained on a large and diverse dataset. This model can then be fine-tuned to recognise additional species relevant to specific regions. “If a client in Europe wants another five species, it will
require the model to be fine-tuned,” Quirion said. “That fine-tuning is a small amount of effort compared to the initial training.”
AI and the future of hydropower
monitoring Quirion sees Species Aware as part of a broader trend towards AI-driven automation in the hydropower sector. “AI is really a way to automate anything that was
previously a manual process,” he said. By automating fish monitoring, AI allows engineers and operators to focus on interpreting results and making informed decisions rather than collecting data. Innovasea ultimately views automated species identification as a way to improve trust – between operators, regulators and the public. “Knowing is everything,” Quirion said. “This technology allows us to know much better and much more accurately about fish around hydro than ever before.” Improved data supports better decision-making,
more effective environmental protection and increased confidence in hydropower as a clean energy source.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45
