Establishment Of An Evaluation Model For Chemical Constituents And Gel Properties Of Silver Carp Surimi Based On Near Infrared Spectroscopy

China is a major producer and processor of aquatic products,with a total output of66,902,900 tons of aquatic products and 21,250,400 tons of aquatic products in 2021,including 1,349,800 tons of surimi products.Surimi is the raw material of surimi products,and its quality directly affects the quality of surimi products.Chemical composition and gel properties are important indexes for evaluation of surimi.The detection of chemical composition indexes mainly adopts the national standard method,while the detection of gel properties requires that surimi be made into gel and then tested.Although the existing detection methods have high accuracy,the experimental process is complicated and the detection cycle is long,which cannot meet the detection requirements of surimi in large-scale industrial production.In recent years,near infrared spectroscopy（NIR）,as a fast,simple and pollution-free analysis technology that can detect multiple indicators at the same time,has developed rapidly in the food industry and has been accepted by more and more factories and enterprises.At present,near infrared spectroscopy is mostly used for the rapid determination of the composition content of surimi,but few studies have reported on the determination of the gel properties of surimi.Based on this,a quantitative evaluation model of water content,crude protein,crude fat,ash,breaking force,breaking distance,water holding capacity and chroma indexes of silver carp surimi was established by using near infrared spectroscopy.Then,based on the predicted values of the near-infrared spectral quantitative models of water,crude protein,crude fat and ash,a stepped-regression model of the breaking force,breaking distance,water holding capacity and chroma of surimi gel was established,and the accuracy of the two modeling methods of surimi gel property index models was compared,in order to realize the rapid detection of the quality of surimi in industrial production and improve the efficiency of surimi production enterprises.The main research contents are as follows:（1）Establishment of NIR quantitative evaluation model for chemical constituents of silver carp.The near infrared spectra of 96 groups of silver carp surimi samples with different contents of water,protein and fat were collected,and the contents of water,crude protein,crude fat and ash were determined by national standard method.The quantitative models of water,crude protein,crude fat and ash of silver carp surimi were established by partial least square method.The best spectral pretreatment methods for moisture,crude protein,crude fat and ash models were moving smoothing+standard normal variable transformation SNV,SG smoothing,mean centralization and detrending,respectively.The characteristic wavelengths selected by CARS were 37,39,53 and 14,respectively.Calibration correlation coefficients Rc were 0.9865,0.9618,0.9586 and 0.9185,respectively.The correlation coefficients Rp were 0.9932,0.9766,0.9798 and 0.9477,respectively.The relative errors SECV of interactive verification are 0.4237,0.4617,0.1698 and 0.0465,respectively.The external test of the model shows that the prediction effect of the model is good and there is no significant difference between the predicted value and the measured value.（2）Establishment of NIR quantitative evaluation model for gel properties of silver carp.96 groups of silver carp surimi samples with near infrared spectral data were prepared into surimi gel,and the indexes of breaking force,breaking distance,water holding capacity and chroma were measured.Partial least square method was used to establish the near infrared spectral quantitative models of breaking force,breaking distance,water holding capacity and chroma of silver carp surimi.The optimal spectral preprocessing methods of breaking force,breaking distance,water holding capacity and chrominance models are normalized,MSC+SG first-order differentiation corrected by multivariate scattering,normalized,SG smoothing+SG first-order differentiation.The characteristic wavelengths selected by CARS were 44,25,25 and 20,respectively.Calibration correlation coefficients Rc were 0.8982,0.6159,0.7098 and 0.6125,respectively.The correlation coefficients Rp were 0.8379,0.4569,0.6636 and 0.7265,respectively.The relative error SECV of interactive verification were 58.9887,0.0848,3.5244 and 1.7069,respectively.The external test of the fitted model shows that the prediction effect of the other models is good except for the breaking distance model,and there is no significant difference between the predicted value and the measured value.（3）Establishment of evaluation model of gel properties of silver carp based o n NIR model of chemical composition.The correlation between the predicted valu e of the chemical composition of silver carp and the measured value of the gel pr operty index was analyzed,and the correlation between the chemical composition i ndex and the gel property index was confirmed.Then the quantitative evaluation model of the gel property of silver carp based on the predicted value of the chem ical composition of near infrared spectrum was established by stepstep regression a nalysis.The indexes of water（₁）,crude protein（₂）,crude fat（₃）and ash（₄）and their square values and cross values were taken as independent variables.Th e breaking force model was₁=2.741X₁²-500.884X₁+102.210X₃²-611.487X₃X₄+22851.091.The breaking distance model is₂=-0.065X₁-0.032X₂-0.905X₄+6.733;The water holding model is₃=-0.211X₁²+31.334X₁-31.804X₄²-1069.720.The chroma model is₄=0.024X₁X₃+77.028.The external test of the fitted model shows that the predict ion effect of the model is good and there is no significant difference between the predicted value and the measured value.The quantitative model of near infrared spectrum of gel properties of silver carp was compared with the average relative error of external test of stepwise regression model.It is concluded that the prediction accuracy of breaking force,breaking distance and chroma model established by stepwise regression analysis is higher,and the prediction accuracy of water holding near infrared spectrum model is higher,and it can be used for rapid detection of surimi in industrial production.