Radmantis, a blue tech company, reports that its underwater AI monitoring system identified fish species composition with 99% accuracy in a hydropower validation study. The study, supported by the U.S. Department of Energy's Water Power Technologies Office, was conducted at Verdantas Flow Labs with rainbow and brown trout. The system detected fish with 95.5% reliability and reconstructed the species mix correctly in 36 of 40 single-blind trials. The technology uses embedded underwater vision and AI inference to classify fish in real time. This could help hydropower operators balance clean energy production with fish passage requirements and environmental compliance. Radmantis is also applying the tech on the Colorado River for invasive species control. The company is seeking partnerships to test the system across more species and field conditions.
