化学计量学
高光谱成像
近红外光谱
计算机科学
工作流程
质量(理念)
遥感
环境科学
过程分析技术
生化工程
质量评定
数据质量
乳品工业
模式识别(心理学)
预处理器
人工智能
成像光谱学
光学(聚焦)
数据科学
偏最小二乘回归
数据挖掘
作者
P. C. Harish Padmanaban,Sathya Bama B
标识
DOI:10.1177/09670335261416167
摘要
As milk adulteration and contamination continue to raise a serious threat to public health, there is a growing demand for rapid, reliable, and non-destructive techniques for milk quality assessment in the dairy industry. This comprehensive review provides recent advancements in non-invasive spectroscopic approaches, with a particular focus on near infrared (NIR) spectroscopy and visible-NIR hyperspectral imaging (Vis-NIR-HSI) for detecting both adulterants and contaminants in milk. Combined with chemometrics and machine learning techniques, these approaches can be used as effective qualitative and quantitative methods. The complete analytical workflow is examined in-depth with the significance of preprocessing strategies on spectral quality, the requirements of wavelength selection, and the roles of targeted and non-targeted chemometric models on classification and regression tasks. Finally, this review addresses existing limitations, including insufficient databases and the complexities of multi-adulterant detection, as well as highlights the emerging trends as future opportunities in multimodal data fusion, Internet of Things (IoT), and deep learning to support real-time monitoring for milk. The purpose of this review is to facilitate the practical, large-scale adoption of robust, non-destructive NIR and Vis-NIR-HSI tools for milk authentication.
科研通智能强力驱动
Strongly Powered by AbleSci AI