Development of an image and spectroscopic predictive model to assess the quality of fishmeal as an ingredient in salmon farm feed

Project title: Development of an image and spectroscopic predictive model to assess the quality of fishmeal as an ingredient in salmon farm feed

Partners: Association of Icelandic Fishmeal Producers, SVN, Eskja, Ísfélagið and the University of Iceland.

Research Fund: Icelandic Food Innovation Fund (is. Matvælasjóður)

Initial year: 2022

Service Category:

Pelagic fish

Contact

Jónas Rúnar Viðarsson

Director of Business and Development

jonas@matis.is

The objective of the project is to develop NIR (near-infrared spectroscopy) prediction model that enables fishmeal producers to get fast, affordable, non-destructive and accurate analysis of the quality of their fishmeal as an ingredient for salmon feed.

Most of the Icelandic fishmeal producers are already using NIR to measure the chemical content of their fishmeal, which are good indicators on the quality of the fishmeal. These parameters are however of limited use when it comes to assessing the quality of the fishmeal for its primary use i.e. as ingredient in aquaculture feed due to the long analytical time for traditional quality assessment. This project will fill in that gap by connecting NIR spectroscopy spectral and hyperspectral imaging data to established quality parameters in salmon farming, in relation to growth and digestibility. This is not a new approach, as Norwegian aquafeed companies have already developed such models, and have been using them to evaluate quality of the fishmeal they source. They do however consider the NIR models as trade secrets, which gives them competitive advantage. Making similar NIR models available for Icelandic fishmeal producers will therefore provide the Icelandic industry with the same (or better) information on the characteristics of their products, as their customers, therefore strengthening their sales and marketing power. No existing prediction models do however exist for the use of hyperspectral imaging data for the mentioned fishmeal quality parameters. The spectral models will also enable fishmeal producers to better assess and improve their own products for internal quality control.

This project follows on an AVS pre-project that finished in 2021 with the publication of a report titled Near-infrared spectroscopy – state of knowledge on the use of NIR in the fishmeal industry. Near-infrared spectroscopy – State of knowledge on the use of NIR in the fishmeal industry.The project will be carried out over two years, where growth and digestibility experiments will be carried out on 20+ types of aquaculture feed in salmon farming in the first year, and in the second year the results will be used to develop the model.