Research On Classification Of Vinegar Based On Electronic Nose System

Vinegar quality affects people's health,there are shoddy phenomenon in the vinegar market caused by diverse brands.The traditional process of detecting vinegar is cumbersome,expensive,long cycle.Detection with electronic nose is faster compared with traditional instruments,but it is not conducive to popularize due to the high price of end products.So,it is of great use to design an electronic nose with the function of detecting vinegar fleetly,conveniently and cheaply.Firstly,this paper introduces the background and development of electronic nose technology,next specifically describes the components of self-made electronic nose and uses it for experiment of data collection.After learning the classical pattern recognition method,a method based on fuzzy Foley-Sammon transform is proposed to classify five kinds of brand vinegar.Finally,the support vector machine(SVM)method was used to classify vinegar as a comparative experiment.Implemented the function of classifying different brands of vinegar rapidly and conveniently by using electronic nose.The main work is as follows:(1)Studied the working principle of electronic nose,of which the hardware and software system are designed.Hardware system mainly includes the sensor array,the power module,the gas sampling device and the data acquisition system.Selected the corresponding sensor on the basis of main components of vinegar;designed the power module according to the requirements of the sensor;the data acquisition card is used to collect data from the consideration of the complexity of implementation;designed the gas sampling device in the light of experimental process.Software includes the host computer program written by LabVIEW,which can transmit the response signal generated by the sensor array to display and store on the PC.(2)In order to be able to classify five kinds of vinegar samples,a method of vinegar classification based on FFST was proposed.It mainly includes several parts of sensor optimization,principal component analysis(PCA),fuzzy Foley-Sammon transform(FFST)and KNN classifier.Sensor optimization provides a "software thinking" method to achieve the purpose of removing data with low contribution rates.Use principal component analysis(PCA)to reduce dimension and computational complexity.The effect of the KNN classifier depends on feature extraction,this paper uses a fuzzy method(FFST)to extract the feature,the set of optimal discriminant vector sets is obtained according to the fuzzy membership degree and the class center,and test sample is projected onto this set to obtain the new sample.The experimental results show that the KNN classifier is used to classify the new samples,and the recognition effect is remarkable.It is further tested about the influence of the K value in the KNN classifier or whether the data is normalized before data analysis on the classification results.The optimal K value is obtained and it is necessary to normalize before vinegar classification.(3)For the purpose of contrast effect,a new method is proposed for vinegar classification based on SVM.First of all,this paper chooses kernel function and the related parameters based on support vector machine theory.And then it describes how to use the software toolkit to implement the algorithm.The experimental results show that the accuracy rate of this method is up to 90.9%,the effect is remarkable.Furthermore,the influence of sensor data optimization and normalization on the performance of SVM model is tested respectively.The results show that the optimization and normalization of the sensor data in the use of the SVM model on the vinegar classification have been improved.In this paper,it is convenient and rapid for self-made electronic nose to classify five different brands of vinegar,classification accuracy is remarkable.And it also can be used for vinegar quality analysis?...