Autonom sortering av stamfisk med integrert bildeanalyse, maskinlæring og genetisk avlsverdier

Offentlig prosjektsammendrag

Highly accurate selection of broodstock is critical for maximising genetic improvement in Atlantic salmon aquaculture. Current methods of selection and sorting of broodstock candidates are based on a mix of multiple step phenotypic assessments, subjective assessments from site staff, and genetic information linked to individual tags. However, more integrated autonomous grading of broodstock using a f data from imaging sensors that will classify a wide suite of growth and welfare traits, together with high resolution genetic data, would deliver a quantum leap in accuracy of broodstock selection, and minimise the negative consequences of fish handling during this process. This project will build on an earlier project that developed an advanced imaging analysis platform for harvest and broodstock, and further develop it as an integrated hardware and software platform that will fully automate sorting of elite broodstock using an optimal mix of phenotypic and genetic data.