Design And Realization Of Salmon Freshness Detection System Based On Electronic Nose Technology

With the continuous improvement of Chinese residents’ living standards and the increasing awareness of healthy life,salmon is favored by people for its rich nutrients.In the market,salmon is usually sold in the way of refrigeration storage,but in this storage environment,due to the biological characteristics of the salmon and the effect of microorganisms,it is easy to cause the spoilage of salmon.Freshness is an important index to evaluate the spoilage and deterioration of fish meat.Traditional fish freshness detection methods include sensory detection,physical and chemical detection and microbial detection,etc.These methods have some problems such as long detection cycle,complicated operation steps and subjective factors.Electronic nose is an artificial olfactory system developed to simulate mammal olfactory,which can quickly detect and identify complex gases with volatile components.Due to the obvious smell change of salmon in the process of spoilage,this paper aims to develop a portable salmon freshness detection system based on electronic nose technology,and provide a fast detection scheme for managers and consumers.The development of this system is divided into three parts: recognition algorithm,hardware design and software design.In recognition algorithm part,support vector machine is used to train salmon freshness prediction model,and the prediction part of algorithm and training model are deployed in electronic nose system to achieve the purpose of rapid distinction of salmon freshness.In the hardware design part,pump suction air path is used as the acquisition device of the system,and appropriate Microelectro Mechanical Systems(MEMS)gas sensor is selected through salmon volatile characteristic gas.Fluid simulation software is used to simulate cuboid air chamber,and the results show that there are problems of air rotation and uneven pressure distribution in cuboid air chamber.According to the structure of human nasal cavity,cuboid air chamber was optimized to solve this problem,and then the location of gas sensor array was determined.The system hardware circuit adopts a modular design and consists of a system core board,a control board and a sensor board.The system core board is responsible for the overall operation and calculation of the system,and the control board provides power,peripheral device interfaces and communication interfaces for the entire system.The sensor board is the Complete data acquisition and signal conversion.In the software design part,based on RT-Thread real-time operation system and Little VGL embedded GUI library,the system human-computer interaction interface,system acquisition,display,storage and other programs are designed,and the upper computer program is designed based on QT platform.After the hardware assembly was completed,the function test was carried out,using ethanol solution as a sample to test the hardware and software functions of the system.Then,the freshness recognition experiment of chilled salmon was carried out.The freshness level of salmon was determined by PH value detection and volatile gas analysis.The model prediction recognition accuracy was 97% through gas characteristic data training,and the trained model was transplanted into electronic nose system to realize real-time detection function.The experimental results show that the system can quickly collect and display the volatile gas data of salmon,and can quickly identify the freshness of salmon,which can provide reference for people to distinguish the freshness of salmon.