摘要
室内有害气体对居民的健康容易产生巨大的威胁。为了提高室内有害气体成分识别的准确率,提出一种基于仿生嗅觉与卷积神经网络相结合的室内有害气体成分识别方法。该方法利用卷积神经网络提取非线性特征的能力对仿生嗅觉的响应信号进行特征提取和成分识别。通过使用气体样本集对该方法进行验证,实验结果表明对不同种类和浓度的有害气体成分识别率均达到88.89%。
Indoor harmful gas is a great threat to the health of residents.In order to improve the accuracy of indoor hazardous gas composition recognition,a novel method based on bionic olfaction and convolutional neural network is proposed.This method uses convolutional neural network to extract nonlinear features and identify components of bionic olfactory response signals.The method was verified by using gas sample set,and the experimental results showed that the identification rate of different kinds and concentra-tions of hazardous gas components reached 88.89%.
作者
潘铭津
何家峰
骆德汉
Pan Mingjin;He Jiafeng;Luo Dehan(School of Informational Engineer,Guangdong University of Technology,Gugangzhou 510000,China)
出处
《信息技术与网络安全》
2019年第12期48-51,共4页
Information Technology and Network Security
基金
广东省省级科技项目(2016A020226018)
关键词
室内有害气体
气体成分识别
仿生嗅觉
卷积神经网络
indoor hazardous gas
gas identification
bionic olfaction
convolutional neural network
