Adopt existing technologies for improved seafood production and to better feed a growing world

Expansion of aquaculture production depends crucially on the development of technologies that are able to perform functions important in a fish farming facility, without human input. For instance, early detection of a disease or parasite outbreak is critical in intensive aquaculture settings. Parasites, such as Sea Lice in Salmonid aquaculture are responsible for large losses. Existing options for dealing with outbreaks are unpalatable. The present Phase I proposal aims to pioneer the use and smooth integration of image detection, machine learning (ML) inference, and automated harvesting, into autonomously operating devices within current aquaculture workflows. Optimized for low-power consumption, our product plans rely on open source frameworks and single board edge computing, with on-board, tensor-flow co-processing. The result will be a flexible and adaptable, automated system that permits the sorting and handling of fish by a number of characteristics, including growth rates/condition, morphological features, behavior, skin abnormalities, diseases such as Bacterial Cold Water Disease (BCWD), or parasite load from sea lice.