Tallinn University of Technology

FishTube – Real-Time AI-Based Underwater Camera System

ABSTRACT 

FishTube is the world's first real-time AI-based underwater camera system, and automatically detects fish in fishways, rivers, lakes and coastal waters. It enhances existing HydroCAM systems with embedded AI to automatically record and label videos only with fish in them, drastically reducing the need for manual processing by human experts.

SUGGESTED USES

  • HYDROPOWER PLANTS: MONITOR FISH MIGRATION AND ENSURE COMPLIANCE WITH ENVIRONMENTAL REGULATIONS INCLUDING THE EU HABITATS AND WATER FRAMEWORK DIRECTIVES.

  • ENVIRONMENTAL AGENCIES: BIODIVERSITY AND FISH BIOMASS MONITORING.

  • RESEARCH ORGANIZATIONS: FISH BEHAVIOR DUIRNG SPAWNING, MIGRATION, HYDROPEAKING, FLOODS AND DROUGHTS.

  • NGOS: VALIDATE THE ECOLOGICAL IMPACTS BEFORE AND AFTER RESTORATION OR DAM REMOVAL USING OUR FULLY AUTOMATED SOLUTION.

ADVANTAGES

  • Efficiency: Filters out unnecessary video content, retaining only relevant data, reducing post-processing by up to 95%.

  • Cost-Effective: Operates at a fraction of the cost of human raters or cloud-based solutions.

  • Real-Time Processing: Uses the latest NVIDIA Orin Nano technology for immediate analysis without cloud reliance.

  • Scalability: Interoperable with various camera systems and adaptable to additional tasks like species recognition.

  • Competitive Edge: First mover in real-time embedded AI for underwater monitoring, no need for cloud-based or manual systems.

ADDITIONAL INFORMATION

State of development 
TRL 7 – The system has been demonstrated in an operational environment and potential customers have tested it, but it is not yet fully market-ready.

Intellectual property info 
Patent and trademark applications submitted.

Media/miscallenous 
https://iamhydro.cam/

Taust