Development Of Electronic Nose Based On STM32 And Its Application In The Identification Of Coffee Flavor

It is of extremely significance in the field of food and agriculture to detect research objects effectively and non-destructively.Traditional detection methods such as sensory or precision instrument evaluation which cannot achieve the purpose of real-time detection rapidly.The sensory detection is not objective for the reason that easily influenced by individual and other factors.Precision instruments is a fair and effective detect method,but the detection process is complicated and expensive.Electronic nose technology is a new method that developed in the 1990’s.With the rapid advancement of science and technology,it has been widely used in agriculture,environment,food and other fields now.Electronic nose is a kind of bionic instrument which simulates biological odor recognition system,it can make qualitative and quantitative analysis of gas over sensor array and pattern recognition method.This paper aims to design an electronic nose system to classify different kinds of gases.The main work contents were as follows:The hardware platform of electronic nose system has been designed and manufactured,which is mainly used for gas signal extraction processing and transmission.Our work including the circuit design with STM32103 MCU as the core,gas sensor selection,sensor array design,gas path design,gas chamber design and so on.The software of electronic nose system had been designed.This paper has accomplished the development of the master computer software and slave computer program.The slave computer program combined with the hardware platform which can extract the gas signal and transmit it to the master computer.The software of the master computer can control the device to collect gas signals as well as display or save the data which transmitted by slave computer.Finally,each module is assembled to form the complete electronic nose system and joins with the pattern recognition method to detect three coffee.The result of experiment showed that the electronic nose system designed in this paper join with the method of principal component analysis(PCA)and Back Propagation neural network are very effective in the classification of coffees.