Research On Nondestructive Detection About Egg Quality Based On Dielectric Properties

Eggs are the important nutrition food,and egg industry is the pillar of Chinese agricultural industry.At present,Chinese egg production is the first in the world.And Chinese per capita occupancy of eggs is third.China has rich resources of eggs.But in the international market share and competitiveness,there is a big gap between China and developed countries,even some parts of the third world countries.This is mainly because technology and means of detection and classification is backward.Therefore,it is the main work of Chinese scientific research workers to study egg quality testing and design testing equipment.For the dielectric property detection technique,the intrinsic relationship between the change of material's electromagnetic properties and the intrinsic component was established to realize the detection of the substance.At present,the detection technology of the dielectric properties of agricultural products has been shown to be superior,and it has been widely used.In this paper,the detection technique of dielectric properties was used for the detection of egg quality.The parallel plate technology and dielectric characteristic testing equipment have been used.The dielectric properties of eggs with different freshness and different varieties were measured.The effects of frequency and freshness on dielectric properties were studied,and the dielectric properties of different varieties of eggs were also studied.A new method based on dielectric properties of egg freshness prediction and variety identification was proposed.Based on the research results,an egg quality detection system based on dielectric properties was developed.The main contents and conclusions of this paper are as follows:(1)In this paper,the effects of frequency and freshness on dielectric properties and dielectric properties of different varieties of eggs were studied.In the test frequency range,the dielectric properties of eggs decreased with the increase of frequency.Under the same test frequency,the relative dielectric constant increased with the freshness index.There were significant differences in the dielectric properties of different varieties of eggs.(2)In this paper,a prediction model of yolk index based on relative dielectric constant was established.The coefficient of determination(R2)of prediction models at different frequencies has been compared.Frequency 1.2 kHz was optimal,and it's coefficient of determination was optimal(R2=0.995).The coefficient of determination between the measured and predicted values was 0.9115,and the error range was±4.2%.(3)In this paper,SVM was used to identify the varieties of eggs.The fusion data of relative dielectric constant and dielectric loss factor was selected as the sample data.The linear kernel function was selected as the kernel function of SVM,and the accuracy rate of such SVM was the highest.Then the SVM parameters were optimized.Two performance indexes were considered,which were optimal working time and classification accuracy.PSO was the optimal selection,and the prediction accuracy in the training set and the test set were both 95.83%.(4)In this paper,an egg quality detection system based on dielectric properties was developed by using Visual Basic 6.0.The egg freshness and variety identification could be achieved by inputting the parameters of egg dielectric properties and basic information to the system.The system was tested to verify the system's completion and detection accuracy.The tested results showed that the prediction error was less than 5%,and the classification accuracy was up to 98%.