Research And Design Of Bionic Olfactory Device For Identification Of Zanthoxylum Bungeanum Varieties

Zanthoxylum bungeanum is an important economic crop in China.It has been planted in large areas in western China such as Sichuan,Shaanxi,Gansu and Qinghai,and has become a pillar industry of the local economy.There are many varieties of Zanthoxylum bungeanum in the market,the shape,color and flavor are different,and the price of different varieties of Zanthoxylum bungeanum is very different.With the increasing emphasis on the deep development and utilization of Zanthoxylum bungeanum,the market demand is growing rapidly,but the phenomenon of inferior filling,aging,adulteration and adulteration in the market has also followed,which has given market supervision and consumers Legal rights have had a serious impact.The existing identification method of Zanthoxylum bungeanum varieties is mainly sensory analysis,however,the accuracy rate of non-professional assessment is very low.In order to solve the problems and realize the identification of Zanthoxylum bungeanum varieties,this paper combined with biomimetic olfactory technology,a scheme for identifying Zanthoxylum bungeanum varieties is proposed,which is a bionic olfactory device.In this paper,SolidWorks,AltiumDesigner and other software were used to design and prototype the bionic olfactory device,and the experimental research was carried out.The main research contents and conclusions are as follows:(1)The design of the system circuit was completed by using AltiumDesigner software.The system consists of overload protection module,temperature control module,fan control module,gas sensor array(composed of 7 gas sensors),data storage module,display module and control chip.The system has the functions of collecting,analyzing and storing the scent data of the Zanthoxylum bungeanum under the condition of constant temperature sealing,and displaying the analysis result on the LCD.(2)The design of the mechanical part of the bionic olfactory device was completed by using SolidWorks software.The design of the mechanical mainly includes the structural design of the closed air chamber,which is the structure and layout design of the gas sensitive detector array,the sample tank,the heating grid,thefan.The designed sealed constant temperature air chamber ensures that the smell of Zanthoxylum bungeanum will not escape,and the constant temperature control of the air chamber is realized,which minimizes the interference of the external environment on the detection.Through the design and layout of the annular sensor array and the directional airway in the air chamber,the spatial positions of the sensors in the sensor array are consistent in the air chamber,and the reliability of the data collected by the sensor array is ensured.(3)The pre-test was carried out by using the produced bionic olfactory device,and the odor data of 157 sets of Zanthoxylum bungeanum samples were collected.The principal component analysis and Wilks statistic were used to reduce the21-dimensional original data set to 6-dimensional,in order to eliminate the useless information in the original data set for the identification of Zanthoxylum bungeanum varieties,and also prepare for the establishment of subsequent models.(4)In order to train the classifier and verify the effect of the two data dimensionality reduction methods,the mathematical model is established by using the original data set,the PCA dimensionality reduction data set and the Wilks statistic dimensionality reduction data set,and the three mathematical models were cross-validated separately.The cross-validation results show that the original correctness rate of the original model based on the original data set is 94.9%,the correctness rate of the A-model which is based on the PCA dimensionality reduction is 96.8%.and the correctness rate of the B-model which is which is based on the Based on the Wilks statistic dimension reduction data set is 96.8%.Therefore,it can be considered that both dimensionality reduction methods improve the reliability of the model.Although the correctness rates of the A model and the B model are the same,but the classifier trained according to the B model has the advantage of being able to directly use the raw data collected by the sensor(unstandardized),which not only improves the efficiency of the classifier,but also reduces the difficulty of writing the classifier programs.Therefore,the classifier is trained according to the B model,and the classifier is transplanted into the control chip of the bionic olfactory device,so that the device has the ability to identify the Zanthoxylum bungeanum variety.(5)In order to test the performance of the bionic olfactory device,30 sets of Zanthoxylum bungeanum samples were selected for the actual test.It was found that it takes about 25 minutes to complete the test of a group of samples(cleaning for about15 minutes,testing and analysis for about 10 minutes).The test results showed that all the 30 groups of samples were correctly identified,and the correct rate reached 100%.The bionic olfactory device designed in this paper basically meets the design requirements and can realize the identification of Zanthoxylum bungeanum varieties.