Research On The Detection Technology Of Beef Flavor Quality And Tenderness Quality Based On Bionic Oral Cavity

Beef is a precious material for people's healthy diet,which can meet people's protein needs,and is a rich source of niacin,vitamin B₆,vitamin B₁₂,phosphorus,potassium,iron and zinc,so it is favored by consumers.Delicious taste,tender and refreshing quality are the evaluation words for beef quality.That is to say,the important indicators of beef quality include the quality of fresh taste and tenderness.However,the freshness and tenderness of beef are usually not evaluated from the first bite,which depends on the feeling of beef when it is transformed in the mouth.Oral cavity is an important place for human beings to perceive the quality of food.The process of food consumption involves various oral operations.Therefore,it is of great significance to detect the taste quality and evaluate the tenderness quality of beef.In this paper,the chewing parameters of chewing process are obtained by sensory evaluation,and then the human oral processing process is simulated based on the bionic oral device.The content of amino acids in beef during chewing was detected by membrane modified sensor,and the release process model was established.The tenderness of beef was detected by the bionic oral device,and the prediction model of tenderness grade was established.At the same time,the tenderness of beef under different chewing parameters was evaluated.Finally,the detection system of beef flavor quality and tenderness quality was established to realize the rapid,accurate,objective and intelligent detection of beef flavor quality and tenderness quality.?1?Research on the detection method of beef flavor amino acids based on bionic oral device.A sensory evaluation group was established through recruitment and training.The group members chewed beef until swallowing.The chewing parameters were recorded and analyzed.According to the data of sensory evaluation,the process of oral processing is simulated based on the bionic oral device.The relationship between the content of amino acids and the electrical signal can be established by using the membrane modified sensor,and the electrical signal can be collected by the electrochemical workstation.The influencing factors of amino acids in beef flavor during mastication were studied.The effects of mastication cycle times,mastication frequency,amount of saliva and bite force on the release of amino acids were analyzed by single factor test and response surface test.By controlling the chewing times,motor speed,peristaltic pump flow and imitating chewing muscle spring of the bionic oral device,the settings of chewing cycle times,chewing frequency,saliva incorporation amount and bite force are realized respectively.The results showed that for the 20 mm×20 mm×10 mm bovine eye muscle,the fitting model of amino acid release?y?and chewing parameters?chewing cycle times X₁,chewing frequency X₂,saliva incorporation X₃,bite force X₄?is as follows:Y=0.086-?1.453E-0.03?X₁+?2.508E-0.03?X₂+?2.882E-0.03?X₃+?8.310E-0.03?X₄-?6.825E-0.03?X₁X₂+?1.025E-0.03?X₁X₃-0.022X₁X₄-0.018X₂X₃-?8.200E-0.03?X₂X₄-0.013X₃X₄+0.013X₁²-?5.556E-0.03?.?2?Construction of the model of amino acid release from beef during chewing.Three models were used to simulate the release of amino acids from beef under the influence of different parameters.Three neural network models are:BP artificial neural network,GA-BP neural network and wavelet neural network.The model was run by MATLAB software,and60 groups of beef were simulated.Determine the parameters of the three models.In order to verify the test results better,the same parameters are selected for the three models.The number of nodes in the input layer is 3,and the number of nodes in the output layer is 1.The hidden layer and other parameters of the three models are determined through continuous operation.The population size of GA-BP neural network is 60,the crossover probability is 0.7,and the mutation probability is 0.01.In order to evaluate the performance of the three neural network models,the predicted value is compared with the actual value,and the prediction accuracy indexes of the three models are compared respectively.The results show that the optimized BP artificial neural network model can effectively predict the content of delicious amino acids in the process of beef chewing.The residual of GA-BP neural network is significantly smaller than the other two models,and the predicted value is closer to the actual value.?3?Research on the detection method of beef tenderness based on bionic oral device.Sensory evaluation and texture analysis represent subjective and objective detection respectively,and the two methods are used to evaluate the tenderness of the same beef.The texture parameters of 60 groups of beef were classified by cluster analysis,and the corresponding grades of beef samples were obtained.The results of sensory evaluation and cluster analysis showed that 55 groups of beef samples were the same.Using biomimetic oral device to convert the force on the occluded beef into electrical signal,the electrical signal collected by the chewing device and the tenderness grade of beef were used to build the model through Fisher discrimination,and 55 groups of beef with the same sensory evaluation and texture detection results were used as the model samples.80%of beef samples were randomly selected for model training,and the remaining 20%were used to verify the accuracy of the model.The accuracy of prediction and discrimination can reach90.9%.Three models were used to evaluate the tenderness of beef under different chewing parameters.The model was run by MATLAB software,and 60 groups of beef were simulated.The parameters of the three neural network models?BP,GA-BP,wavelet?are the same.The input layer of the model is the number of chewing cycles,chewing frequency,the amount of saliva,and the output layer is the tenderness level of beef.Compare the accuracy of the three network models to select the optimal model,and analyze the predicted value of the three models.The results show that the accuracy of the three neural network models are 85%,95%and 90%,respectively.So GA-BP model can be used to evaluate the tenderness of beef during chewing.?4?The development of the detection system for beef taste quality and tenderness quality.Based on the software MATLAB 2017b and the software SQL Server 2016,a set of software system for rapid detection of beef flavor quality and tenderness quality is developed,so as to achieve accurate and rapid detection of beef quality.The interface of each module is designed by GUI.The expert system of beef taste quality and tenderness quality detection consists of seven modules,which are user login,sample taste detection,fresh amino acid signal acquisition,fresh response value output,sample tenderness detection,tenderness quality signal acquisition and tenderness quality result output.Using the beef sample information database module built in SQL Server 2016 in the system,the data of beef sample has been well managed and can read and write database data.The software system has the functions of parameter setting,control and operation of bionic oral device,real-time display of detection response potential value,collection of detection data,curve drawing of response value results,grade prediction of beef tenderness quality,data saving of analysis results,etc.The development of the detection system of fresh and tender quality makes the detection process more simple and efficient.