电子鼻的深度神经网络算法实现可燃气体检测的研究  

Research on Combustible Gas Detection Based on Deep Neural Network Algorithm of Electronic Nose

作　　者：房瑞山 薛莹莹 陈畅明 向奕 王平[1,2] 万浩[1,2] FANG Ruishan;XUE Yingying;CHEN Changming;XIANG Yi;WANG Ping;WAN Hao(Biosensor National Special Laboratory,College of Biomedical Engineering and Instrument Science,Zhejinag University,Hangzhou Zhejiang 310027,China;Key Laboratory for Biomedical Engineering of Education Ministry,College of Biomedical Engineering and Instrument Science,Zhejiang University,Hangzhou Zhejiang 310027,China)

机构地区：[1]浙江大学生仪学院生物传感器国家专业实验室,浙江杭州310027 [2]浙江大学生仪学院生物医学工程教育部重点实验室,浙江杭州310027

出　　处：《传感技术学报》2024年第9期1518-1524,共7页Chinese Journal of Sensors and Actuators

基　　金：国家重点研发计划课题项目(2021YFC3300303)。

摘　　要：提出了一种将格拉姆角场与深度残差卷积神经网络相结合的电子鼻气体识别算法,可以提高燃气中常见混合危险气体的识别准确度。通过对电子鼻的气体样本数据都进行格拉姆角场(Graham Angle Field,GAF)变换,使得上位机接收到的二维传感器响应数据,经过升维后成为可以输入到卷积神经网络(Convolutional Neural Network,CNN)中的三维数据形式,从而发挥了CNN特征提取能力强、模型收敛快和识别准确率高的优势。实验结果表明,该算法在干扰气体存在下的情况下,对CO和CH4的检测准确率分别达到93.04%和92.43%,通过与多种主流算法进行对比和分析,表明该算法具有抗干扰性高、检测准确率高的优点,为实际环境中干扰气体存在时的可燃气体高可靠性和特异性检测提供了一种具有良好应用前景的智能识别算法。A gas identification algorithm for electronic noses that combines the Graham angle field with a deep residual convolutional neural network is proposed,which can improve the identification accuracy of common mixed dangerous components in gas.Graham angle field(GAF)transformation is performed on the gas sample data of the electronic nose,so that the two-dimensional sensor response data received by the host computer can be input to the convolutional neural network(CNN)after dimension enhancement in the form of three-dimensional data,thus giving full play to the advantages of CNNs strong feature extraction ability,fast model convergence and high rec-ognition accuracy.The experimental results show that in the presence of interfering gases,the detection accuracy of the algorithm for CO and CH4 reaches 93.04%and 92.43%,respectively.Compared with the conventional principal component analysis,linear difference analysis and support vector machine identification methods,the proposed algorithm has the advantages of high anti-interference and high detection accuracy,and provides an intelligent identification algorithm with good application prospects for the high reliable and specific detection of combustible gases in the actual environment when interfering gases exist.

关 键 词：电子鼻 气体识别 模式识别 卷积神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP212.9[自动化与计算机技术—控制科学与工程]