基于轻量级神经网络MobileNetV3-large的黄茶闷黄程度判别  

Identification of yellowing degree of yellow tea based on lightweight neural network MobileNetV3-large

作　　者：葛炳钢 张旭雯 刘岁 杨亚[1,2] 周铁军 傅冬和[1,2,3] GE Binggang;ZHANG Xuwen;LIU Sui;YANG Ya;ZHOU Tiejun;FU Donghe(Key Lab of Tea Science of Ministry of Education,Changsha,Hunan 410128,China;National Research Center ofEngineering and Technology for Utilization of Botanical Functional Ingredients,Changsha,Hunan 410128,China;Co-Innovation Center for Provincial-Ministry(Education Ministry)for Utilization of Botanical Functional Ingredients,Changsha,Hunan 410128,China;College of Information and Intelligence,Hunan Agricultural University,Changsha,Hunan 410128,China)

机构地区：[1]茶学教育部重点实验室,湖南长沙410128 [2]国家植物功能成分利用工程技术研究中心,湖南长沙410128 [3]植物功能成分利用省部(教育部)共建协同创新中心,湖南长沙410128 [4]湖南农业大学信息与智能科学技术学院,湖南长沙410128

出　　处：《湖南农业大学学报（自然科学版）》2024年第1期91-99,共9页Journal of Hunan Agricultural University(Natural Sciences)

基　　金：湖南省创新型省份建设专项(2021NK1020);国家“十四五”重点研发计划项目(2022YFD2101102);湖南省科技重点研发项目(2018NK2035);湖南农业大学研究生科研创新项目(2022XC064)。

摘　　要：以碧香早品种为材料,通过相机采集不同闷黄时长下的闷黄叶图像共675张,建立了3种闷黄程度的黄茶样本数据集,采用位置变换、随机亮度、增加对比度、添加噪声、随机缩放操作对闷黄叶图像集进行数据增强,运用迁移学习方法,在ImageNet数据集取得MobileNetV3-Large的预训练模型,对迁移网络的所有权重信息进行训练,最终建立了针对黄茶闷黄程度的轻量级卷积神经网络MobileNetV3-Large识别模型,并利用Grad-CAM热力图可视化和置信分数监控黄茶品质的变化。结果表明:经训练后的MobileNetV3-Large模型测试的识别准确率达到98.51%,精确率为99.10%,召回率为98.93%,加权分数为98.20%;MobileNetV3-Large模型的识别准确率高于传统机器学习模型SVM、XGBoost和KNN;通过Grad-CAM热力图可视化显示,MobileNetV3-Large模型在不同的识别场景下能够准确定位并提取闷黄叶特征,准确地识别闷黄程度。可见,MobileNetV3-Large模型有较好的泛化性,可以快速、无损地识别黄茶的闷黄程度。In this study,three dataset of yellow tea sample data were collected from a total of 675 Bixiangzao variety tea images of yellowing leaves at different stage.Position transformation,random brightness,increasing contrast,adding noise,and random scaling operations were used to enhance the data.The pre-training model of MobileNetV3-Large was obtained from the ImageNet data sets using transfer learning method,and ownership weight information of the network was transferred for training,then the lightweight convolutional neural network MobileNetV3-Large recognition model for yellowing degree of yellow tea was established.Grad-CAM(Gradient-weighted class activation mapping)heat map visualization and confidence scores were used to monitor the changes in yellow tea quality.The results of using the model to estimate the teas showed that the trained MobileNetV3-Large model achieves recognition accuracy of 98.51%,precision of 99.10%,recall of 98.93%and F1-score of 98.20%.The recognition accuracy of MobileNetV3-Large model was higher than the traditional machine learning models SVM(Support vector machine),XGBoost(eXtreme gradient boosting)and KNN(K-Nearest neighbors).As visualized by Grad-CAM heat map,the MobileNetV3-Large model was able to accurately locate and extract yellowing leaf features in different recognition scenarios,and accurately identify the yellowing degree.In conclustion,the MobileNetV3-Large model had better generalization and could rapidly and nondestructively identify the yellowing degree of yellow tea.

关 键 词：黄茶 闷黄程度识别 卷积神经网络 迁移学习 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]