Non‐destructive detection of milk nutritional components based on hyperspectral imaging DOI Open Access

Yuanpu Zhang,

Jiangping Liu

Journal of Food Science, Год журнала: 2024, Номер unknown

Опубликована: Дек. 28, 2024

Abstract As consumers increasingly prioritize food safety and nutritional value, the dairy industry faces a pressing need for rapid accurate methods to detect essential components in milk, such as fat, protein, lactose. Hyperspectral imaging (HSI) technology, known its non‐destructive, fast, precise nature, shows great promise quality assessment. However, high dimensionality of HSI data poses challenges effective band selection model optimization. Additionally, prior studies primarily focus on predicting single without addressing simultaneous multi‐component detection. To overcome these challenges, this study presents comprehensive approach that integrates moving average smoothing first derivative (MA‐FD) preprocessing, improved coati optimization algorithm (ICOA), CatBoost multi‐target regression. ICOA incorporates good point set strategy, dynamic opposition‐based learning, golden sine algorithm, which significantly enhance global search capability convergence speed selection. Combined with CatBoost's prediction capability, method enables detection lactose levels milk. Experimental results demonstrate accuracy, calibration achieving an coefficient determination (MultiR 2 ) 0.9992 root mean square error (MultiRMSE) 0.0240, while yielded MultiR 0.9797 MultiRMSE 0.1181. Prediction R values were 0.9658, 0.9910, 0.9825, respectively. The proposed demonstrates robust predictive accuracy reliability milk assessment, potential application broader assessments is substantial. Practical Application This provides rapid, non‐destructive assessing by detecting key through hyperspectral imaging, combined MA‐FD selection, offers reliable, non‐invasive solution supports control helps safeguard consumer health.

Язык: Английский

Hyperspectral estimation of soil organic matter using improved spotted hyena optimizer and iteratively retained informative variables DOI
Hui Zhang, Yunbo Shen, Huanhuan Lv

и другие.

Microchemical Journal, Год журнала: 2025, Номер unknown, С. 113410 - 113410

Опубликована: Март 1, 2025

Язык: Английский

Процитировано

0
Hyperspectral Imaging and Deep Learning for Quality and Safety Inspection of Fruits and Vegetables: A Review DOI
Yang Chen, Zhiming Guo, Douglas Fernandes Barbin

и другие.

Journal of Agricultural and Food Chemistry, Год журнала: 2025, Номер unknown

Опубликована: Апрель 16, 2025

Quality inspection of fruits and vegetables linked to food safety monitoring quality control. Traditional chemical analysis physical measurement techniques are reliable, they also time-consuming, costly, susceptible environmental sample changes. Hyperspectral imaging technology combined with deep learning methods can effectively overcome these problems. Compared human evaluation, automated improves efficiency, reduces subjective error, promotes the intelligent precise fruit vegetable inspection. This paper reviews reports on application hyperspectral in various aspects assessment. In addition, latest applications technologies fields safety, internal quality, external reviewed, challenges future development directions this field prospected. has shown significant advantages inspection, especially improving accuracy efficiency. Future research should focus reducing costs, optimizing equipment, personalizing feature extraction, model generalizability. lightweight models balance accuracy, enhancement database importance quantitative be brought attention. These efforts will promote wide improve its practicability actual production environment, bring important progress for management.

Язык: Английский

Процитировано

0
Non‐destructive detection of milk nutritional components based on hyperspectral imaging DOI Open Access

Yuanpu Zhang,

Jiangping Liu

Journal of Food Science, Год журнала: 2024, Номер unknown

Опубликована: Дек. 28, 2024

Abstract As consumers increasingly prioritize food safety and nutritional value, the dairy industry faces a pressing need for rapid accurate methods to detect essential components in milk, such as fat, protein, lactose. Hyperspectral imaging (HSI) technology, known its non‐destructive, fast, precise nature, shows great promise quality assessment. However, high dimensionality of HSI data poses challenges effective band selection model optimization. Additionally, prior studies primarily focus on predicting single without addressing simultaneous multi‐component detection. To overcome these challenges, this study presents comprehensive approach that integrates moving average smoothing first derivative (MA‐FD) preprocessing, improved coati optimization algorithm (ICOA), CatBoost multi‐target regression. ICOA incorporates good point set strategy, dynamic opposition‐based learning, golden sine algorithm, which significantly enhance global search capability convergence speed selection. Combined with CatBoost's prediction capability, method enables detection lactose levels milk. Experimental results demonstrate accuracy, calibration achieving an coefficient determination (MultiR 2 ) 0.9992 root mean square error (MultiRMSE) 0.0240, while yielded MultiR 0.9797 MultiRMSE 0.1181. Prediction R values were 0.9658, 0.9910, 0.9825, respectively. The proposed demonstrates robust predictive accuracy reliability milk assessment, potential application broader assessments is substantial. Practical Application This provides rapid, non‐destructive assessing by detecting key through hyperspectral imaging, combined MA‐FD selection, offers reliable, non‐invasive solution supports control helps safeguard consumer health.

Язык: Английский

Процитировано

0