Research On Food Freshness Detection Based On Electronic Nose And Sensory Evaluation

With the improvement of living standards,people pay more attention to food quality,and food safety.The conventional detection methods for food freshness are usually cumbersome,time consuming,and vulnerable to subjective interference.Therefore,it is necessary to find a simple,fast,scientific and objective method.Electronic nose(E-nose)detection technology has the advantages of portability,nondestructiveness,and real-time.With the continuous development of material science and computer technology,E-nose research has developed rapidly.It has great application value to identify the freshness of food by using E-nose in daily life.Therefore,a food freshness detection and recognition system based on electronic nose is designed and developed in this thesis.The major research contents and innovations of this thesis are as follows:1.The differences in food volatile components of different types of food under different sensory freshness levels is compared,and the gas markers related to spoilage food is initially determined.Three kinds of foods--meat,vegetables and fruits--are analyzed by solid phase micro extraction-gas chromatography-mass spectrometry(SPME-GC-MS)by collecting the airbag headspace gas in the food spoilage process.The difference in volatile compounds of three kinds of foods under different sensory freshness is measured,and the correlation between changes in food sensory characteristics and volatile components is confirmed.Then,based on the experimental results,the gas markers related to food spoilage were preliminarily determined: ethanol,ammonia,and hydrogen sulfide,and these three gases were used as the main target gases for system detection.2.The sensor array is screened and calibrated by target gases.And an electronic nose system for food freshness detection based on MOS gas sensors is constructed according to the design requirements of modularization and constructed according to the design requirements of modularization and miniaturization,The metal oxide semiconductor(MOS)gas sensor is selected and the gas sensor array is designed according to the actual food spoilage related characteristic gas.And the electronic nose detection system is divided into airway module,detection and control module,and system software module according to the modular method.The designed electronic nose system can detect and collect food odor information and transmit data through the system software.It has the advantages of small size,low cost,and easy integration with other equipment.3.The electronic nose system was used to detect and recrd the corruption of12 kinds of common food under separate storage,and a linear discrimination algorithm is used to develop a single food freshness discrimination model.The model is tested with actual data.12 kinds of foods from meat,vegetables and fruit that are common in daily life are selected as experimental objects,and the human sensory evaluation is use as a reference to establish sensory freshness evaluation standards for different foods.The food freshness is divided into three types: freshness,sub-freshness and spoilage category.Then the electronic nose system is used to test 12 kinds of food,recording the sensor response changes of different foods from fresh to spoiled under separate storage condition.The actual test data is used to establish a linear discriminant analysis model to evaluate the food freshness.The results show that,the linear discriminant analysis algorithm can effectively distinguish the different freshness levels of the tested food under separate storage,and the recognition accuracy of different food freshness is at least 85.77%,and the best is 100.00%.4.In view of the complex storage conditions of mixed foods,corresponding experiments are designed,and the electronic nose system is used to test the mixed storage food samples.The food freshness recognition model is established based on the convolutional neural network and is compared with other algorithms.In view of the complex situation of food mixing and storage in daily life,the mixed food testing experiment was designed by restricting the types of experiment food and simplifying the mixing method,and the electronic nose system is used to test certain food mixing conditions.The mixed food identification results are divided into four categories: freshness,meat spoilaged,vegetable spoilaged and fruit spoilaged,that is,further judgments are made on the type of spoiled food on the basis of judging whether the food sample contains spoiled food.The application of convolutional neural network(CNN)is explored in the freshness recognition of mixed foods.The algorithm is compared to 3 traditional pattern recognition methods,and the results prove that the accuracy of convolutional neural network can reach 93.33%,which is the best among the tested algorithms.The results show that the recognition model based on CNN can make effective judgments on the freshness of mixed foods,and has better performance and great application potential.