Rapid estimation of fat content in salmon fillets by colour image analysis

The aim of this study was to develop a simple method for the automatic estimation of fat content in salmon fillets by means of image analysis. Salmon fillets have a zebra-like appearance, consisting of white myocommata stripes divided by red-coloured muscle tissue. The stripes contain a high proportion of lipid, and it has previously been shown that the proportion of myocommata in a fillet correlates with its fat content. A possible measurement method for fillet fat content might therefore be to use image analysis to determine the area of the white stripes visible on the fillets surface compared to the total area of the fillet. Fifteen salmon fillets were sampled from an assembly line at a local fish-processing plant, photographed and analysed for lipid. The results obtained by the image analysis were compared with those from chemical analysis, and showed a good correlation ( r = 0.84 ). Although the sample size was relatively small, the correlation was high enough to suggest that the method, after development and improvement, could prove useful as a simple tool in salmon aquaculture research and product processing.