基于气味检测的茶叶种类识别方法研究
Research on tea category identification method based on odor detection
杨志蒙 1黄波 1赵永礼1
作者信息
- 1. 上海工程技术大学 机械与汽车工程学院,上海 201620
- 折叠
摘要
提出一种利用电子鼻系统检测茶香气味辨别茶叶种类的识别方法,使用主成分分析(PCA)、K-means聚类和卷积神经网络(CNN)3 种机器学习方法对 10 种茶叶种类进行识别.实验结果表明,基于PCA降维特征的K聚类精度为 85.17%,比基于原始特征的K聚类精度 78.83%更优,基于PCA降维特征的CNN算法识别率达到 95%,基于茶香气味检测的茶叶种类识别方法方便快捷,具有可行性.
Abstract
An identification method that uses an electronic nose system to detect the tea aroma to identify the tea categories was proposes.Three machine learning methods were used to identify 10 types of tea,namely principal component analysis(PCA),K-means clustering,and convolutional neural network(CNN),respectively.The experimental results showed that the accuracy of K-clustering based on PCA dimensionality reduction features was 85.17%,better than the accuracy of K-clustering based on original features of 78.83%.The identification accuracy of the CNN algorithm based on PCA dimensionality reduction features could reach to 95%.It confirms that the tea category recognition method based on tea aroma and odor detection is efficient and feasible.
关键词
茶叶种类/气味识别/电子鼻系统/机器学习Key words
type of tea/odor recognition/E-nose system/machine learning引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
维普
万方数据