首页|基于电子鼻和电子舌与1D-CNN-LSTM模型的花椒产地快速溯源检测

基于电子鼻和电子舌与1D-CNN-LSTM模型的花椒产地快速溯源检测

扫码查看
原文链接
  • 万方数据
  • 维普
针对不同产地花椒产品的溯源问题,提出一种基于电子鼻和电子舌结合一维卷积神经网络(One Dimension-Convolutional Neural Networks,1D-CNN)-长短期记忆网络(Long Short-Term Memory,LSTM)混合模型的花椒产地快速检测方法.以5个不同产地的花椒为试验对象,采用电子舌和电子鼻分别采集花椒样本的味觉和嗅觉指纹图谱信息,根据信号特点分别设计1D-CNN提取味觉和嗅觉信号中的局部空间特征,然后采用LSTM捕捉信号的时间序列特征,最后采用多层感知机融合两种特征并进行分类识别.实验结果表明,电子鼻与电子舌信息融合对不同产地花椒的分辨准确率优于单一设备,与其他深度模型相比,所提的模型分类准确性更高,其准确率、精确率、召回率、F1分数分别达到99.0％、99.1％、99.0％、0.989.以上研究将为不同产地花椒的快速鉴定提供新的方法,并为其他农产品的产地溯源检测提供新的研究思路.
Rapid Traceability Detection of Zanthoxylum bungeanum Origin Based on Electronic Nose and Electronic Tongue with 1D-CNN-LSTM Model
For the rapid traceability detection of Zanthoxylum bungeanum products of different origins,a hybrid one dimension-convolution-al neural networks(1D-CNN)-long short-term memory(LSTM)model based on electronic tongue and electronic nose is proposed. Taking Zanthoxylum bungeanum from five different habitats as the test samples,the electronic tongue and nose are used to collect the taste and smell fingerprint information respectively. According to the signal characteristics,1D-CNN is designed to extract the local spatial features of taste and smell signals,and LSTM is used to capture the time series features. Finally,multi-layer perceptron is used to fuse the two fea-tures and distinguish the categories. The experiment results show that the combination of electronic nose and electronic tongue data can distinguish Zanthoxylum bungeanum of different origins with better accuracy than a single equipment. The proposed model has higher clas-sification accuracy compared to other depth models with the accuracy,precision,recall and F1-score of 99.0％,99.1％,99.0％,0.989 re-spectively. This will provide a new method for the rapid identification of Zanthoxylum bungeanum from different origins,and a new re-search idea for the traceability detection of other agricultural products.

sensor signal processingzanthoxylum bungeanumorigin traceabilityelectronic noseelectronic tonguefeature fusionconvolution neural networklong short-term memory

张擎、杨晓婧、金鑫宁、陈立同、高文、王志强、姜春磊

展开 >

山东理工大学计算机科学与技术学院,山东 淄博 255000

淄博市工业数字经济发展中心,山东 淄博 255000

烟台黄金职业学院信息工程系,山东 烟台 265401

传感器信号处理 花椒 产地溯源 电子鼻 电子舌 特征融合 卷积神经网络 长短期记忆网络

山东省自然科学基金项目教育部科技发展中心产学研创新基金项目

ZR2019MF0242018A02010

2024

10.3969/j.issn.1004-1699.2024.05.023

传感技术学报
东南大学 中国微米纳米技术学会

传感技术学报

CSTPCD北大核心
影响因子：1.276
ISSN：1004-1699
年,卷(期)：2024.37(5)