Research On Detection Method For Food Safety Based On Terahertz

Food safety is a momentous problem related to people's lives and national security.A large number of potential food safety hazards are caused by the long-term storage of food in realistic produce.Research shows that with the increase of food storage time,the nutrients contained in the food will be largely lost,and the microorganisms in the food will also rapidly propagate.Therefore,food safety detection technology has attracted extensive attention from researchers and scientists.In recent years,terahertz technology has made great progress,and many studies have introduced it into the field of food safety detection,which has become a very promising method.In the master degree thesis,the change of the terahertz wave's spectral parameters propagating in apples with different storage time was studied.Then according to this change,the storage days of apples were classified by machine learning algorithm.This method provides the theoretical and experimental bases for exploring a food storage time detection technology based on terahertz waves.In the master degree thesis,the classification method of food storage time based on terahertz technology is studied and analyzed,which mainly includes four aspects: food scattering parameter collection,establishing scattering parameter model,calculating permittivity and using permittivity to classify storage time.The main contributions of the master degree thesis can be summarized as follows:(1)The theory of food storage time detection based on terahertz technology is discussed.The changes in the water content of food during storage were analyzed.Terahertz waves are very sensitive to water.Therefore foods with different water content will have different strengths of absorption to terahertz waves,which explains how terahertz technology detects food storage time.(2)Taking the apple as an example,models of food scattering parameters and frequencies in the terahertz band are established.A vector network analyzer was used to measure apple's scattering parameters in the 0.75-1.1THz band,and the relationship between scattering parameters and frequency was modeled using the ridge regression model and support vector regression model.The data obtained by the model effectively reduces the noise,and can better reflect the relationship between the scattering parameters and the frequency.(3)The method of calculating the permittivity through scattering parameters is studied.For the problem that the permittivity is not easy to be measured directly through experiments,the relationship between the scattering parameters and the permittivity is derived based on the tangential boundary conditions of the electromagnetic field,and then the permittivity of the apple in the terahertz frequency band is calculated.(4)The classification method of food storage duration based on the permittivity of terahertz frequency band is explored.Based on the calculated permittivity,apple's storage days were classified using traditional machine learning algorithms and neural network algorithms,respectively.The results validate the previous theoretical analysis that using the change of the terahertz band permittivity in different storage times of food,the storage time can be effectively detected.In summary,the terahertz-based food storage time classification method studied in this master degree thesis provides an idea for food quality inspection,and also shows the application potential of terahertz in the field of food safety inspection.