Study On Electronic Nose Gas Sensing Method Based On Light Absorption

Electronic nose?e-nose?is an electronic technology that simulates the function of biological olfactory organs for gas identification.In an e-nose,gas sensor array,signal preprocessing unit and pattern recognition unit are used to realize olfactory perception,which can avoid human error with good repeatability.However,limited by its core unit?gas sensor array?,e-nose also has some defects,such as limited detection range caused by small sensor array,long response/recovery time,easy aging/poisoning,short life and so on.As an important means of gas detection,optical absorption gas sensing technology has following advantages:wide detection range,fast response,and the sensor unit has strong response specificity and broad spectrum.These characteristics meet the requirements of e-nose for its sensor array.Therefore,it has great potential to introduce light absorption gas sensing technology into e-nose.Aiming at the defects of narrow detection range and long sensor response/recovery time of traditional e-nose,a gas sensing method of broadband light absorption e-nose?optical e-nose?based on existing researches was studied further in this thesis which completed the actual construction and validation of optical e-nose sensing platform.In order to solve the problem of limited spectral resolution in optical e-nose,space heterodyne spectroscopy technology was introduced to optimize the optical e-nose?an improved optical e-nose?.Based on the above targets,specific research contents are as follows:?1?Optical electronic nose gas sensing methodAn absorbance detection mechanism and corresponding gas sensing model of optical e-nose were constructed according to the Lambert-Beer law,and an experimental platform was built by high-performance grating spectrometer and other devices.Then,different kinds of gases with different concentrations were selected for experiments,and one-dimensional absorbance data of gases were derived as sensing data.Finally,the sensing data were analyzed by signal preprocessing and pattern recognition algorithm.Results show that recognition accuracies of different gas response data are more than92%,the response time is 30 seconds and the platform can detect dozens of gases,which verifies the effectiveness of the optical e-nose gas sensing method.?2?Optimization of optical electronic nose gas sensing methodLimited by grating spectrum detection technology,the spectral resolution of the optical e-nose is not high enough,which limits the detection performance of the optical e-nose.To solve this problem,spatial heterodyne spectroscopy with wide spectrum and ultra-high resolution was introduced to optimize the optical e-nose.Then a transmission ratio spectrum detection mechanism and corresponding gas sensing model were established,and an experimental platform was built.Response images of different bands were obtained as sensing data.According to characteristics of the image,an image feature extraction algorithm based on directional filter bank?DFB?was proposed,and the sensing data were mapped to feature space to realize data filtering.Finally,pattern recognition algorithms were used to judge the feature data.Results show that the proposed DFB algorithm improves the average recognition rate of the test set from 74%to 97%,and spatial heterodyne technology improves the spectral resolution from1.5mm^-1 to about 0.1mm^-1,which verifies the superiority of the DFB algorithm and the effectiveness of the optical e-nose optimization method.