In the fast-paced world of modern agriculture, ensuring the quality of fresh produce during storage and transportation is a persistent challenge. Traditional detection methods, while accurate, are often destructive, time-consuming, and require specialized expertise. Enter a groundbreaking study published in *智慧农业*, which introduces an embedded fluorescence imaging detection system that promises to revolutionize the way we monitor fruit and vegetable quality.
The research, led by Chenhong Gao, Qibing Zhu, and Min Huang from the School of Internet of Things Engineering at Jiangnan University, addresses critical bottlenecks in current detection technologies. “Fresh fruits and vegetables are prone to quality deterioration due to microbial proliferation and changes in enzyme activity,” explains Gao. “While advanced optical detection technologies like hyperspectral imaging offer non-destructive advantages, they are often expensive, bulky, and lack portability. Our goal was to develop a low-cost, efficient solution that meets the demands of the modern supply chain.”
The team’s innovative system integrates fluorescence imaging technology with an embedded platform, utilizing a lightweight deep learning model based on YOLOv8. The system employs a 365 nm, 10 W ultraviolet LED as the excitation light source and a CMOS camera for image acquisition. The real magic, however, lies in the algorithmic improvements. By replacing the original YOLOv8 backbone network with MobileNetV4 and applying channel pruning techniques, the researchers significantly reduced the computational load while maintaining high accuracy.
“Our improved YOLOv8-MobileNetV4 model achieved a mean average precision of 95.91% for three-level quality classification,” says Zhu. “This model not only outperformed other mainstream lightweight models in terms of accuracy but also exhibited faster detection speeds, making it an excellent balance between precision and efficiency.”
The practical implications for the agriculture sector are substantial. The system’s low hardware cost, compact size, and portability make it an ideal solution for real-time, non-destructive quality monitoring in fruit and vegetable supply chains. This technology could streamline operations, reduce waste, and enhance food safety, ultimately benefiting both producers and consumers.
Looking ahead, the researchers plan to expand the sample library to include more produce types and mixed deterioration levels. They also aim to further optimize the algorithm to improve robustness in complex multi-target scenarios. “Our ultimate goal is to provide a comprehensive, reliable tool for quality detection that can be seamlessly integrated into existing agricultural practices,” Huang adds.
As the agriculture industry continues to evolve, innovations like this embedded fluorescence imaging detection system will play a pivotal role in shaping the future of food quality assurance. By combining cutting-edge technology with practical solutions, this research paves the way for more efficient, sustainable, and safe agricultural practices.