.png){kind=icon}
.png){kind=icon}
Return
Development of a Nano‑functionalized Biopolymer Film Using Deep Learning for Real‑Time Relative Humidity Monitoring in Fresh Food Packaging
.png){kind=icon}
PRE
AI
F
D
Q
B
L
DOI：10.1007/s11947-026-04501-z.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
An intelligent relative humidity (RH)-sensing label was developed by integrating a nano-functionalized chitosan/polyvinyl alcohol (PVA) biocomposite with a deep learning (DL) model for real-time monitoring in fresh food packaging. Nanocomposite films were synthesized by incorporating zinc oxide nanoparticles (ZnO-NPs) or titanium dioxide nanoparticles (TiO2-NPs) into a chitosan/PVA matrix with methylene blue as an indicator. Comprehensive characterization confirmed successful nanoparticle integration, which formed stable interactions with the polymer network as confirmed by FTIR, XRD, and UV–Vis analyses. The ZnO-NPs samples demonstrated optimal performance, exhibiting superior sensitivity, stability, and a controlled hydrophilic balance (water contact angle: 49.8°) compared to the TiO2-NPs samples. The label delivered a distinct colorimetric response (green → blue → black) to elevated RH (> 75% RH). Five DL architectures were evaluated for classifying RH levels from the label’s colorimetric response. The Xception model emerged as the best-performing, demonstrating superior training accuracy (95.2%) and robust generalization ability (95.65% overall accuracy). The practical efficacy of integrating the nano-enhanced label and the Xception DL model was conclusively validated through real-time monitoring of RH within packaged silver carp and apples. This work presents a promising, automated solution for intelligent RH monitoring in fresh food supply chains.
Keywords:
Relative humidity sensing
Biopolymer composite film
Nanocomposites
Deep learning models
Food packaging
Journal
IF:
5.8
Papers:
4.5K
Citations:
1.5W
Organization
Cited Papers