Innovative integration of computer vision, IoT, and digital twin in food quality and safety assessment

文献类型： 外文期刊

第一作者： Guo, Mengshuai

作者： Guo, Mengshuai;Lv, Xin;Wang, Dan;Chen, Hong;Wei, Fang;Wei, Fang

作者机构：

关键词： Food quality and safety; Deep learning; Computer vision; Digital twin; IoT

期刊名称：TRENDS IN FOOD SCIENCE & TECHNOLOGY （ 影响因子：15.4； 五年影响因子：18.4 ）

ISSN： 0924-2244

年卷期： 2025 年 163 卷

页码：

收录情况： SCI

摘要： Background: Ensuring food quality and safety is a key priority for public health and economic stability. Traditional methods of food quality assessment, while effective, are often labor-intensive, destructive or lack traceability and transparency. Recent advances in deep learning and computer vision introduce digitally intelligent, cost-effective and automated solutions. Scope and approach: This review presents a typical workflow of deep learning and computer vision, from data acquisition and data preprocessing to model selection, training and evaluation for validation, and summarizes the applications of deep learning and computer vision in different areas of food, such as image classification, object detection, image segmentation, and image generation, as well as model optimization strategies for different tasks. The applications of Internet of Things (IoT), digital twin, computer vision, and deep learning technologies in the food industry are highlighted. In addition, this review also discusses transfer learning and model compression methods, and reviews the applications of lightweight models and embedded systems in the food industry. Key findings and conclusions: The innovative integration of technologies such as computer vision, deep learning, IoT, and digital twin has enhanced food traceability and transparency, and promoted sustainable development. The advancement of cloud computing and big data technologies has promoted the deep integration of these technologies, enabling real-time, accurate and dynamic decision-making in food production. Looking forward to the future, the focus of future research should be placed on improving the availability and quality of labeled datasets, enhancing the interpretability and robustness of model.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序