Study On Shelf Life Prediction Model Of Freshness And Quality Of Salmon

The rapid non-destructive testing of quality shelf life of aquatic products from production to processing is a topic of concern.However,traditional quality inspection methods are time consuming,destructive,and complex.Therefore,the fish industry urgently needs a fast and accurate prediction method to simultaneously measure the amount of chemical and microbial spoilage under variable temperature conditions.In this study,salmon was used as the research object.Determination of freshness,freshness rating and shelf life of salmon stored at-2,0,4,10 °C.The superiority of RBFNNs model was determined and the model support was provided by comparing the Arrhenius shelf life prediction model with the RBFNNs.The volatile gas composition and content of salmon at various storage stages at-2,0,4 and 10 °C were determined by electronic nose.The prediction model of PCA-RBFNNs salmon freshness based on electronic nose was constructed.The computer vision system was used to collect the pupil color space values of different storage stages stored at 0 °C,and a computer vision system based on salmon pupil color parameters combined with multiple regression models was constructed..The main findings were as follows:1.The upper limit of the acceptable freshness of salmon fillets at shelf life of-2,0,4 and 10 °C was 21 d,12d,10 d and 72 h respectively;TAC,TVB-N,K and hist amine values Changes can better reflect the freshness changes of salmon fillets,pr oviding a basis for subsequent experiments and prediction models.;2.With the shelf life stored at-2,0,4,and 10 °C as the comparison target,the relative error between the predicted and experimental values of the two models: the relative error of RBFNNs shelf-life prediction model is within ±2%(except the 2 day),and the relative error of Arrhenius shelf-life prediction model fluctuates between ± 22.63%-±1.35%greatly.So,the prediction accuracy of the RBFNNs shelf-life prediction model is higher than the Arrhenius.3.Ammonia/amines,hydrocarbons,organic solvents and aromatic compounds were detected by the electronic nose during storage.GC-IMS analysis confirmed the change in gas phase composition.The prediction models were established by RBFNNs and PCA the based on electronic nose.The relative errors of TBA,TVB-N,TAC,K and histamine values in PCA-RBFNNs model were within ±10%,and SA was within ±15%.These results indicate that the PCA-RBFNNs model based on electronic nose,which can be used to predict changes in the freshness of salmon fillets at-2 °C to 10 °C;4.The three sets of color spaces indicate that the color parameters associated with the change in freshness of the salmon were R,G,B,L*,I value better.Relevant color parameters combined with various freshness indicators.A freshness prediction multiple regression model based on pupil color parameters was constructed.The model had good predictive effect on simultaneous determination of TBA,TVB-N,TAC,K value and histamine value.The correlation coefficients R2 were 0.9997,0.9989,0.9992,0.9995,0.9995 and the F values were 589.42,186.26,259.82,428.72,and 436.54,respectively.It indicated that the freshness prediction regression model is significant.