AI Fish Species Identification Hits 99% Accuracy in Hydropower Study

A U.S. Department of Energy-supported validation study found that Radmantis' underwater AI monitoring system can identify fish species composition with 99% accuracy and detect fish with 95.5% reliability in dense, fast-moving conditions. The two-month test at Verdantas Flow Labs used rainbow and brown trout, which look similar, to stress the system's ability to distinguish between species. The platform correctly reconstructed the exact mix in 36 of 40 single-blind trials and stayed within five fish in the rest. Hydropower operators need to balance clean energy production with fish passage and ecosystem health. This technology provides continuous, real-time species-level data that could improve compliance with environmental regulations, support native species recovery, and help control invasive species. Radmantis is already applying similar tech on the Colorado River and is looking for partners to expand field testing across more species and operating conditions.