Smartphone-based pH titration for liquid food applications

IF 2.2 4区 化学 Q2 Engineering
Yuhui Xiao, Yaqiu Huang, Junhong Qiu, Honghao Cai, Hui Ni
{"title":"Smartphone-based pH titration for liquid food applications","authors":"Yuhui Xiao,&nbsp;Yaqiu Huang,&nbsp;Junhong Qiu,&nbsp;Honghao Cai,&nbsp;Hui Ni","doi":"10.1007/s11696-024-03715-9","DOIUrl":null,"url":null,"abstract":"<div><p>The pH detection helps control food quality, prevent spoilage, determine storage methods, and monitor additive levels. In the previous studies, colorimetric pH detection involved manual capture of target regions and classification of acid–base categories, leading to time-consuming processes. Additionally, some researchers relied solely on <i>R*G*B*</i> or <i>H*S*V*</i> to build regression models, potentially limiting their generalizability and robustness. To address the limitations, this study proposed a colorimetric method that combines pH paper, smartphone, computer vision, and machine learning for fast and precise pH detection. Advantages of the computer vision model YOLOv5 include its ability to quickly capture the target region of the pH paper and automatically categorize it as either acidic or basic. Subsequently, recursive feature elimination was applied to filter out irrelevant features from the <i>R*G*B*</i>, <i>H*S*V*</i>, <i>L*a*b*</i>, Gray, <i>X</i><sub><i>R</i></sub>, <i>X</i><sub><i>G</i></sub>, and <i>X</i><sub><i>B</i></sub>. Finally, the support vector regression was used to develop the regression model for pH value prediction. YOLOv5 demonstrated exceptional performance with mean average precision of 0.995, classification accuracy of 100%, and detection time of 4.9 ms. The pH prediction model achieved a mean absolute error (MAE) of 0.023 for acidity and 0.061 for alkalinity, signifying a notable advancement compared to the MAE range of 0.03–0.46 observed in the previous studies. The proposed approach shows potential in improving the dependability and effectiveness of pH detection, specifically in resource-constrained scenarios.</p><h3>Graphical abstract</h3><div><figure><div><div><picture><source><img></source></picture></div></div></figure></div></div>","PeriodicalId":513,"journal":{"name":"Chemical Papers","volume":null,"pages":null},"PeriodicalIF":2.2000,"publicationDate":"2024-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Chemical Papers","FirstCategoryId":"92","ListUrlMain":"https://link.springer.com/article/10.1007/s11696-024-03715-9","RegionNum":4,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"Engineering","Score":null,"Total":0}
引用次数: 0

Abstract

The pH detection helps control food quality, prevent spoilage, determine storage methods, and monitor additive levels. In the previous studies, colorimetric pH detection involved manual capture of target regions and classification of acid–base categories, leading to time-consuming processes. Additionally, some researchers relied solely on R*G*B* or H*S*V* to build regression models, potentially limiting their generalizability and robustness. To address the limitations, this study proposed a colorimetric method that combines pH paper, smartphone, computer vision, and machine learning for fast and precise pH detection. Advantages of the computer vision model YOLOv5 include its ability to quickly capture the target region of the pH paper and automatically categorize it as either acidic or basic. Subsequently, recursive feature elimination was applied to filter out irrelevant features from the R*G*B*, H*S*V*, L*a*b*, Gray, XR, XG, and XB. Finally, the support vector regression was used to develop the regression model for pH value prediction. YOLOv5 demonstrated exceptional performance with mean average precision of 0.995, classification accuracy of 100%, and detection time of 4.9 ms. The pH prediction model achieved a mean absolute error (MAE) of 0.023 for acidity and 0.061 for alkalinity, signifying a notable advancement compared to the MAE range of 0.03–0.46 observed in the previous studies. The proposed approach shows potential in improving the dependability and effectiveness of pH detection, specifically in resource-constrained scenarios.

Graphical abstract

基于智能手机的液态食品 pH 滴定仪
pH 值检测有助于控制食品质量、防止变质、确定储存方法和监控添加剂水平。在以往的研究中,比色法 pH 值检测需要人工捕捉目标区域并对酸碱类别进行分类,耗费大量时间。此外,一些研究人员仅仅依靠 R*G*B* 或 H*S*V* 来建立回归模型,这可能会限制其通用性和稳健性。针对上述局限性,本研究提出了一种结合 pH 纸、智能手机、计算机视觉和机器学习的比色法,以实现快速、精确的 pH 检测。计算机视觉模型 YOLOv5 的优点包括能够快速捕捉 pH 纸上的目标区域,并自动将其分为酸性或碱性。随后,采用递归特征消除法过滤掉 R*G*B*、H*S*V*、L*a*b*、Gray、XR、XG 和 XB 中的无关特征。最后,使用支持向量回归法建立 pH 值预测回归模型。YOLOv5 的平均精度为 0.995,分类准确率为 100%,检测时间为 4.9 毫秒,表现出卓越的性能。酸碱度预测模型的平均绝对误差(MAE)为 0.023(酸度)和 0.061(碱度),与之前研究中观察到的 0.03-0.46 的 MAE 范围相比有了显著提高。所提出的方法在提高 pH 值检测的可靠性和有效性方面显示出潜力,特别是在资源有限的情况下。 图表摘要
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Chemical Papers
Chemical Papers Chemical Engineering-General Chemical Engineering
CiteScore
3.30
自引率
4.50%
发文量
590
期刊介绍: Chemical Papers is a peer-reviewed, international journal devoted to basic and applied chemical research. It has a broad scope covering the chemical sciences, but favors interdisciplinary research and studies that bring chemistry together with other disciplines.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信