基于气味检测的红富士苹果新鲜度识别方法研究  

Research on the freshness recognition method of red Fuji apples based on odor detection

作　　者：杨明丽 纠海峰 邓薇 Yang Mingli;Jiu Haifeng;Deng Wei(School of Light Industry,Harbin University of Commerce,Harbin 150028,Ch)

机构地区：[1]哈尔滨商业大学轻工学院,哈尔滨150028

出　　处：《国外电子测量技术》2024年第10期91-101,共11页Foreign Electronic Measurement Technology

基　　金：哈尔滨商业大学青年创新人才培育计划(XL0106)项目资助。

摘　　要：食品安全是人民健康的基石,为实现对苹果新鲜度的快速、无损检测,提出一种应用电子鼻系统检测红富士苹果新鲜度的识别方法,该电子鼻系统是由数据采集系统和模式识别系统两部分构成,分别实现对苹果气味的浓度数据采集及数据分析。利用主成分分析、K-means聚类、反向传播神经网络和鲸鱼算法优化的反向传播神经网络多种学习算法对采集数据进行分析。实验结果表明,基于主成分分析法降维特征的K-means聚类算法识别准确率为70.67%,其值优于基于原始特征的Kmeans聚类算法识别率。利用鲸鱼算法优化的反向传播神经网络识别精度可达到95%,可以准确识别不同新鲜度的样本,该识别方法高效、快速、便捷,可行性较强。Food safety is the cornerstone of people's health.To achieve rapid and non-destructive detection of apple freshness,a recognition method for detecting the freshness of Red Fuji apples using an electronic nose system is proposed.The electronic nose system consists of two parts:A data acquisition system and a pattern recognition system,which respectively collect and analyze concentration data of apple odor.Using various learning algorithms such as principal component analysis,K-means clustering,back propagation neural network,and whale algorithm optimized back propagation neural network to analyze the collected data.The experimental results show that the recognition accuracy of K-means clustering algorithm based on principal component analysis for dimensionality reduction features is 70.67%,which is better than the recognition rate of K-means clustering algorithm based on original features.The recognition accuracy of back propagation neural network optimized by whale optimization algorithm can reach 95%,which can accurately identify samples with different freshness levels.This recognition method is efficient,fast,convenient,and highly feasible.

关 键 词：苹果新鲜度 电子鼻系统 气味识别 学习算法 

分 类 号：TN307[电子电信—物理电子学]