Cel This is a feasibility study on the inline defect detection of fish fillets by image analysis. A feasibility study was made of inline defect detection in fish fillets using image analysis.The preliminary results of the study of defect detection in fillets of plaice from the fish processing industry show that detection and distinction of blood defects and black membranes is possible under laboratory (static) conditions.One of the important fish processing unit operations is filleting. After filleting the fish is trimmed manually for defects like bones, blood spots, remnants of skin and black membranes and textural faults. The trimming operation is laborious and it is impossible to guarantee fillets free from defects. Image analysis for detection for fish fillets is expected to be an extremely good tool for automation of grading, selecting and sorting. Image of plaice fillets, with different types of defects like bones, blood spots and remnants of skin, will be recorded by different cameras and different sources of illumination with different wavelengths. Appropriate filters will be used. The images will be analysed and appropriate software for detecting the fish fillet defects will be developed for grading, selecting and sorting. A prototype of an inline image detection system for fish fillet defects will be designed and tested on a semiindustrial scale. Dziedzina nauki natural sciencescomputer and information sciencessoftwareengineering and technologyelectrical engineering, electronic engineering, information engineeringelectronic engineeringsensorsoptical sensorssocial sciencessociologyindustrial relationsautomation Program(-y) FP2-FAR - Community research and coordination programmes (EEC) in the fisheries sector, 1988-1992 Temat(-y) Data not available Zaproszenie do składania wniosków Data not available System finansowania Data not available Koordynator TNO/IVP-Afdel. Inst. voor Visserij Producten Wkład UE Brak danych Adres Dokweg 37, Postbus 183 1970 AD Ijmuiden Niderlandy Zobacz na mapie Koszt całkowity Brak danych