Research And Application Of Quality Detection Algorithm In Pork Cold Chain Transportation

People’s demand for fresh pork and other fresh products is becoming more and more intense.When buying pork,people pay more attention to the quality of pork.The traditional evaluation of pork freshness mainly relies on artificial recognition,including observing color,smelling odor,pointing and so on.However,these methods are only practical for some experienced talents,and for most people,they cannot accurately judge the freshness of pork by these methods.At the same time,the emergence of COVID-19 has made some contact methods to measure pork freshness even less feasible.People are looking forward to a more rapid,non-contact and high accuracy method for pork freshness detection.Under the influence of "African Swine Fever",the cold chain transportation of pork will replace the transportation of live pigs as the main circulation mode.At present,most of the pork on the market is cold meat.Take Beijing as an example.The consumption of pork in Beijing is very large and the self-sufficiency rate is low,requiring a large amount of imports from foreign ports.Pork transportation involves cold chain logistics.Whether pork cold chain logistics is qualified is related to the stability of Beijing pork market supply,which is an important topic related to people’s "vegetable basket".In this paper,the freshness classification task of chilled pork is realized by designing fresh experiment of pork cold chain transportation combined with deep learning related technologies,and the results are applied to pork cold chain transportation to ensure the food safety of residents in the capital.The key parts of this article are as follows:(1)This paper introduced the research background and significance of chilled pork freshness classification,and studied the research status and theoretical basis of pork freshness classification and deep learning at home and abroad..(2)The sample data set of chilled pork freshness was established.Through experimental design,photos of chilled pork with different freshness in cold chain transportation environment were collected.Then,a variety of image enhancement methods are used to expand the dataset so that it can meet the requirements of deep learning.(3)I have studied the classic models of convolutional neural networks such as VGG16,Inception and Resnet-50 in depth.Experiments were carried out on the pork data set using the network model and relevant parameters were optimized for comparison.Through comparative study,the advantages of each network model are explored,and then I build a new convolutional neural network model by combining these models.The classification accuracy of 97%is achieved through training.(4)I deployed the trained model to the server to develop the pork freshness recognition system.The computer terminal pork freshness monitoring system in the process of cold chain logistics is realized,and the mini apps for identifying pork freshness on the mobile terminal is also developed.