With digital innovative power into the future of aquaculture

With the project "Digitalisation of land-based shrimp farming in Germany using AI-based evaluation of imaging & acoustic systems", shrimp farming is being made fit for the future. 

While industries such as media, food, telecommunications and banking have already come a long way in terms of digitalisation, aquaculture is still in its infancy, according to a report by the Boston Consulting Group. Yet digitalisation could bring decisive progress, especially in the area of animal health; since the fish farmer does not know the exact biomass of fish/shrimp in his tanks (but only the calculated biomass), overfeeding can occur if the animals grow less than expected, for example. This is not only detrimental to the fish farmer's yield, but can also have a negative impact on animal welfare or health. This is because the feed residues remaining in the system lower the water quality and increase bacterial growth. Feed resudies also impair the function of the filter technology and lead to inaccurate values in the measuring probes used to control the water quality.

The problem arises because in many aquaculture operations the amount of feed is measured on the basis of theoretically calculated feeding tables and not on the actual feeding behaviour of the animals. Often the turbidity of the water makes it difficult for the staff to manually adjust the feed quantities to the feeding activity of the animals in order to prevent overfeeding.

With the MonitorShrimp project, shrimp feeding in the future will no longer be based on imprecise feeding tables but on precisely calculated biomass: Modern image and sound recording technologies in combination with artificial intelligence, developed among others by the project partner MonitorFish GmbH, already provide the basis for optimised digital feeding. For example, feeding behaviour can be recorded and evaluated indirectly via acoustic sounds or typical movement patterns. "What we lack above all is biological data," says Dominik Ewald, Managing Director of MonitorFish. The project partners AWI, Fördegarnelen GmbH and Erwin Sander Elektroapparatebau GmbH are now working on this data for the first time, both on a laboratory scale and on the farm. Using specially developed filters and computer programmes, it will be possible to record images animals, even when water is turbid. Ultimately, this allows calculate the biomass in all tanks using digital size recording on the basis of annotations and artificial learning programmes.

After two years, the project partners plan to bring a software to market that will enable shrimp farmers to feed more efficiently in the future, ultimately produce more economically and with optimised animal welfare.

The project will run for two years. The project partners are MonitorFish GmbH, Erwin Sander Elektroapparatebau GmbH and Förde Garnelen GmbH & Co. KG. The project is coordinated by the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI).

Funded by the Federal Agency for Agriculture and Food (BLE), the project "Digitalisation of land-based shrimp farming in Germany using AI-based evaluation of imaging & acoustic systems" was officially launched on 04.02.

News

Test it!

A first demo version can be tested here. On this page you can upload your own picture of your shrimp and have their number counted. If you don't have any pictures and maybe not even a tank with shrimps, you may use one of our photos (photo 1, photo 2, photo 3, photo 4). The demonstration version does not currently work with high counting accuracy under all farm conditions. For example, different exposures, animal sizes or other smartphone types still reduce the counting accuracy. These factors are to be accounted for before the end of the project period in order to obtain a counting accuracy of >95% under different farm conditions.

New article about MonitorShrimp in ínternational Aquafeed magazine (december 2021/ page 42)

Internship - Bachelor thesis - Master thesis

Internships and final theses can be carried out in this project.

 

Partner