融合声振信号与可见近红外透射光谱的苹果轻度霉心病检测  

Detection of Mild Moldy-Core Disease in Apples by Fusing Acoustic-Vibration Signals and Visible-Near-Infrared Transmission Spectroscopy

作　　者：谷家辉 赖丽思 王凯[1] 张慧[1] GU Jiahui;LAI Lisi;WANG Kai;ZHANG Hui(College of Intelligent Manufacturing Modern Industry,Xinjiang University,Ürümqi 830017,China)

机构地区：[1]新疆大学智能制造现代产业学院,新疆乌鲁木齐830017

出　　处：《食品科学》2024年第23期259-267,共9页Food Science

基　　金：新疆维吾尔自治区自然科学基金项目(2022D01C674)。

摘　　要：针对单一方法对苹果轻度霉心病检测精度较低的问题,提出基于近红外透射光谱和声振技术的异源信息融合方法,以提升对苹果轻度霉心病的判别能力。针对近红外光谱信号,首先分析不同预处理和特征提取方法对建模效果的影响,完成光谱特征波段的选择。针对声振信号,利用YSV工程测试与信号分析软件和Pearson相关系数优选7个时域特征。随后,通过特征拼接将光谱特征波段与时域特征组成融合特征向量,分别采用卷积神经网络(convolutional neural networks,CNN)、长短时记忆网络(long short-term memory,LSTM)和CNN-LSTM基于单一源特征和融合特征构建判别模型。通过模型性能分析,融合了近红外透射光谱15个特征波段与7个时域特征的CNN-LSTM组合模型对于轻度霉心病的判别性能最优,测试集的准确率、召回率、特异性和F1分数分别达到了98.31%、97.06%、97.06%和97.90%。实验结果证明本研究提出的可见近红外透射光谱与声振信号特征融合方法可以有效提高苹果轻度霉心病的判别准确率。In response to the issue of low detection accuracy for mild moldy-core disease in apples using single methods,this study proposed an approach based on the fusion of near-infrared transmission spectroscopy and acoustic-vibration technology to enhance the discriminability of mild moldy-core disease in apples.For the near-infrared spectral signals,the impacts of different preprocessing and feature extraction methods on modeling outcomes were analyzed to select the spectral feature bands.For the acoustic-vibration signals,7 time-domain features were optimized by using the YSV engineering test and signal analysis software as well as calculating Pearson correlation coefficients.The spectral feature bands and time-domain features were then concatenated to form a fused feature vector.Convolutional neural networks(CNN),long short-term memory(LSTM),and CNN-LSTM were employed to construct discrimination models based on single and fused features,separately.The performance analysis of the models revealed that the CNN-LSTM combination model,which integrated 15 near-infrared transmission spectral bands and 7 time-domain features,exhibited the best performance in discriminating mild moldy-core disease,with accuracy,recall,specificity,and F1 scores of 98.31%,97.06%,97.06%,and 97.90%on the test set,respectively.These findings demonstrate that the proposed method can effectively improve the discrimination accuracy of mild moldy-core disease in apples.

关 键 词：可见近红外透射光谱 声振信号 苹果霉心病 特征融合 卷积神经网络-长短时记忆网络 

分 类 号：TS207.7[轻工技术与工程—食品科学]