Project
Automatic shrimp triage (NL SORTERING)
In the research project ‘Accurate selection” ("De Juiste Sortering" in Dutch) aims at innovative processing of shrimp catches on board shrimp fishing vessels. Image recognition is a crucial new element in the new machine. The innovation partners are machine builder De Boer RVS, ILVO, and skipper Jan-Jurie van Eekelen of the WR 9. In the processing line, the detection line must be able to triage the catch into marketable shrimp, undersized shrimp and other bycatch on the basis of camera images and automatic image recognition. Shrimp smaller than 6.8 mm and bycatch go back into the sea as quickly as possible. Marketable shrimp are further subdivided into three market classes that are stored separately on board.
Research approach
The research project consists of several components: Exchanging the developed knowledge with relevant stakeholders at home and abroad and finding and refining the best adjustments so that the camera detection line makes the right separation between size and undersize shrimp. We evaluate in each development phase the selection accuracy of the innovative processing system compared to the accuracy in a traditional processing line (rinse sorter and exit sieve). We assess the (differences in ) survival rate of bycatch, when using the innovative processing line compared to the survival rate of bycatch in the traditional processing line. We also train the system on automatic recognition of bycatch species.
Relevance/Valorisation
Four benefits are expected to result from this new sorting: a more accurate sorting of the catch, a higher survival rate of bycatch, improved working conditions for fishermen and the possibility for more and cheaper data collection for fisheries research. The researchers are convinced that the co-creative nature of this development, and the combined scientific and practical expertise, greatly increases the likelihood of success.