Study On Nondestructive Testing Methods Of Preserved Egg Gel Quality

Preserved eggs are a kind of traditional pickled egg products in China.They are made by pickling fresh eggs in lye.At present,according to the factory inspection requirements,artificial detection methods are generally used to test the gel quality of preserved eggs.The preserved eggs are divided into three types(high-quality eggs,inferior eggs and bad eggs)according to the gel quality,and the eggs are knocked by hand.According to the elasticity of the preserved eggs,the preserved eggs are divided into elastic eggs and non-elastic eggs.Good elasticity also means good gel quality.Manual testing has weaknesses such as high labor intensity,low efficiency.and strong subjectivity.It is necessary to find non-destructive testing methods to detect and classify the quality of preserved eggs,so as to provide technical support for the automatic detection of the quality of preserved eggs.The main research contents and conclusions are as follows:(1)The relationship between the texture parameters of preserved eggs and gel quality was studied.Through the texture profile analysis test,the springness,hardness,cohesiveness,cohesion,and chewiness properties of two kinds of gel quality preserved eggs(high-quality eggs and inferior eggs)are analyzed to compare the quality of different eggs.The texture parameters of the eggs are better than those of the inferior eggs.The SPSS software was used to analyze the correlation between texture parameters and preserved egg gel quality grade,and the correlation between the two was obtained.(2)Established a visual non-destructive testing model for the quality of preserved eggs gel.A preserved egg image acquisition platform was built based on machine vision technology to collect the preserved eggs transmission image,preprocess the preserving egg image,remove the image background,and extract R,G,B,H,S,V,L,a,b,?R,?G,?B,?H,?S,?V,?L,?a,and ?b.Principal component analysis was performed on this 18 feature parameters,and the obtained principal components were input into the model for training.The results show that three types of gel-quality preserved eggs(high-quality eggs,inferior eggs and bad eggs)cannot be classified based on machine vision technology,but they can be divided into edible eggs(high-quality eggs,inferior eggs)and inedible eggs(bad eggs).The best classification model is GA-SVM compared with different model classifications.The recognition rate of the test set is 97.56%,and the recognition rate of bad eggs is 100%.(3)Established a spectral detection model for the quality of preserved eggs gel and a comprehensive visual-spectral detection model.Based on the near-infraredspectroscopy technology to collect spectral data of three types of gel-quality preserved eggs(high-quality eggs,inferior eggs and bad eggs),it was found that the use of near-infrared spectroscopy alone could not classify the three types of gel-quality preserved eggs,but it could classify edible eggs(high-quality eggs,inferior eggs)that cannot be distinguished by machine vision technology.Therefore,the original spectral data of high-quality eggs and inferior eggs are pre-processed by multiple scattering correction,and CARS is used to reduce the spectral data to extract their characteristic wavelengths and establish a classification model.The results show that near-infrared spectroscopy can classify high-quality eggs and inferior eggs.The recognition rate of high-quality eggs is 96.49%,and the recognition rate of inferior eggs is 94.12%.Therefore,the distribution detection method is proposed to classify the three types of gel-quality preserved eggs.The machine vision technology is used to classify the preserved eggs into edible eggs(high-quality eggs,inferior eggs)and inedible eggs(bad eggs),and then use near-infrared spectroscopy technology.The edible eggs(high-quality eggs and inferior eggs)were divided into high-quality eggs and inferior eggs,and the overall recognition rate was 96.38%.(4)Established an elastic vibration non-destructive testing model for the quality of preserved egg gel.Use the acceleration sensor to build a vibration signal acquisition platform for preserved eggs,collect the vibration signals of the preserved eggs after being excited,and convert the time domain data of the vibration signals into frequency domain data by fast Fourier transform,and extract their peak and average values,variance,root mean square,maximum value,minimum value,square root value,kurtosis,waveform factor,margin factor,pulse factor,and peak factor are the characteristic parameters in the time domain,and the main response frequency,maximum amplitude,and center of gravity frequency are extracted in the frequency domain,The mean square frequency and the root mean square frequency are characteristic parameters.Principal component analysis was performed on the characteristic parameters to reduce the data dimension,and a model for the classification of preserved eggs was established.Comparing the classification results of different models,the GA-SVM model has the best classification effect,with an overall recognition rate of 85.53% and inelastic eggs(inferior eggs)recognition rate of94.59%.It shows that the acceleration sensor can be used to classify the preserved egg's elasticity.