Font Size: a A A

Research And Design Of Intelligent Tibetan Tea Grade Identification Device Based On Simulated Olfactory Technology

Posted on:2021-11-21Degree:MasterType:Thesis
Country:ChinaCandidate:X YangFull Text:PDF
GTID:2481306506959269Subject:Master of Agriculture
Abstract/Summary:PDF Full Text Request
Ya'an Tibetan tea belongs to six types of black tea.Black tea in China was produced in Sichuan.Ya'an is the starting point of Sichuan-Tibet Road,the ancient tea-horse road [1].As a characteristic drinking tea,Tibetan tea has a very high status in the Ya'an area.Due to its own caffeine content and process,it not only has the particularity of not affecting sleep quality,but also has the function of reducing fat and losing weight [2].With the development of the Ya'an region in recent years,the Ya'an Tibetan tea market has gradually expanded,and various types of Ya'an Tibetan tea in the market have also gradually increased.Due to its special craftsmanship and different rating methods from other teas,it has brought new The challenge has had a serious impact on the legitimate rights and interests of consumers.The existing method for identifying the grade of Tibetan tea is sensory identification,and the probability of non-professionals independently recognizing its grade is extremely low.To solve the above problems and realize the popularization of the classification of Tibetan tea,this paper combines machine learning,bionic smell and other technologies A method for identifying the grade of Tibetan tea by a machine,namely an intelligent Tibetan tea grade identification device,is proposed.This paper uses Soild Works,MATLAB,Altium Designer and other software to design a bionic olfactory device,and conducted experimental research.The main research contents and conclusions are as follows:(1)Design the system circuit through Altium Designer software.The system consists of control chip and peripheral control circuit,data acquisition module,display module,data storage module,temperature control module,fan module and protection circuit module.The system realizes the data collection,analysis and storage of the smell of Tibetan tea under the conditions of constant temperature,closed and circulating air channels,and displays the analysis results on the display screen to identify the grade of Tibetan tea.(2)Use Solid Works software to complete the mechanical structure design of the system acquisition device.The mechanical structure design part requires the construction of a closed air chamber,and includes the structural design of the circulating air duct that is easy to clean,that is,the structure and layout of the sensor array for data acquisition,sample tank,heating device,fan,and circuit placement.The airtight thermostatic air chamber is designed to ensure that under constant temperature,the smell of Tibetan tea quickly fills the air chamber,and will not be emitted,as far as possible to reduce the interference of the external environment on the detection.The design structure of the circulating air duct reduces the interval of each data collection,and can quickly complete the cleaning for the next set of tests.The sensor array used for data collection is placed in the designed groove,and the sensor probe is placed at the airway entrance to ensure that each sensor can be evenly sampled in the environment of circulating wind,ensuring the reliability of the data.(3)Use the system involved to carry out data collection tests.The test samples include 4grades of Tibetan tea,which are Kangzhu special grade,Kangzhu ordinary,Jinjian special grade and Jinjian ordinary.A total of 280 sets of Tibetan tea data were collected,and 70 sets of Tibetan tea odor data for each grade.In the MATLAB environment,using deep neural network,Fisher discriminated to model the four kinds of collected tea gas data,the results show that the deep neural network test set identification accuracy rate is the highest,reaching97.5%.The trained model is implanted into the control chip to complete the intelligent collection of tea grade identification device that integrates data collection,analysis and discrimination.After using the device,the four types of Tibetan tea were re-verified 10 times,and the verification accuracy rate reached 95%,indicating that the Tibetan tea grade identification device designed in this paper basically met the design requirements and could achieve the identification of Tibetan tea grade.
Keywords/Search Tags:Ya'an Tibetan tea, grade identification, deep neural network, bionic olfactory device
PDF Full Text Request
Related items