A Hyperspectral-Based Method For The Detection Of Umami Substances And Umami Intensity In Siniperca Chuatsi

The umami intensity is an important index to evaluate the quality of freshwater fish meat.The existing detection methods are highly subjective and time-consuming,and there is an urgent need to study reliable and efficient non-destructive methods for umami intensity detection.In this study,the meat of Siniperca chuatsi from different days and origins was used as the research object,and the detection model of umami substances and umami intensity of the meat of Siniperca chuatsi was established by combining with hyperspectral technology,and the main findings were as follows:(1)A hyperspectral detection model of umami amino acids in the meat of Siniperca chuatsi was determined.The full-wavelength detection model and simplified model of the umami amino acids were established with different pretreatment methods,different modeling methods and different feature wavelength screening algorithms,respectively,using the meat pieces of Siniperca chuatsi and minced fish as the research objects.In order to optimize the hyperspectral-based umami substance detection model,this paper proposes to optimize the support vector machine(SVM)modeling parameters C and g by using social network search(Sns)algorithm and the golden jackal(GJ)algorithm to select the feature The Golden jackal(GJ)algorithm was proposed to optimize the algorithm for selecting the combination of wavelengths and to establish a simplified model for the detection of umami amino acids.The results showed that the optimal full-wavelength model for Glu content in the meat of Siniperca chuatsi was the model of Sns-SVR with SG(Savitzky Golay Convolution Smoothing)preprocessing of fish block spectra,and its root mean square error of prediction set RMSE_P=0.0831 mg/100 g.The optimal simplified model is the GJ-Sns-SVR model,and the root mean square error of the model is RMSE_P=0.0600mg/100g when 20 optimal feature wavelengths are selected.The optimal simplified model was the GJ-Sns-SVR model,and the number of combinations of the 25optimal feature wavelengths was selected as 25,and the root mean square error of the model was RMSE_P=0.0806mg/100g.(2)A hyperspectral detection model for the umami nucleotides of Siniperca chuatsi meat was determined.The results showed that the optimal full-wavelength model for AMP content in the meat of Siniperca chuatsi was the Sns-SVR model with SNV(Standard Normal Variate Transformation)preprocessing of the fish block spectra,and the root mean square error of its prediction set was RMSE_P=0.1100 mg/100 g.The optimal simplified model is the GJ-Sns-SVR model,and the root mean square error of the model is RMSE_P=0.1179mg/100g when 37 optimal characteristic wavelengths are selected.The optimal full-wavelength model for the IMP content in the flesh of Siniperca chuatsi is the Sns-SVR model with SNV preprocessing of the fish block spectra,and the root mean square error of the prediction set is RMSE_P=0.1418mg/100g.The optimal simplified model is the GJ-Sns-SVR model,and the root mean square error of the model RMSE_P=0.1578mg/100g when 25 optimal feature wavelengths are selected.(3)The detection model of hyperspectral application to the umami intensity of Siniperca chuatsi meat was determined.The full-wavelength detection model and the simplified model for the umami intensity of Siniperca chuatsi meat were developed by combining different pretreatment methods,modeling methods and different feature wavelength screening algorithms.The results showed that the optimal full-wavelength model for the umam intensity in the meat of Siniperca chuatsi was the Sns-SVR model based on the original spectrum,with the root mean square error of the prediction set RMSE_P=0.1245.The optimal simplified model was the GJ-Sns-SVR model,and the root mean square error of the model prediction set RMSE_P=0.1257 when 17 optimal feature wavelengths were selected.The results of this study show that hyperspectral-based detection of the umami substance and umami intensity of Siniperca chuatsi can be achieved,which provides a new idea for the detection of the umami intensity of fish and a theoretical basis for the development of an instrument for the detection of the umami intensity of fish.