Automatic recognition of snack packaging time information based on YOLOv11

Abstract

Food safety and information transparency are closely related to national economy and people's livelihood. Against the backdrop of bagged and boxed food dominating the market, accurate and clear labeling and efficient identification of production date and shelf life on packaging are of great significance. However, manual quality inspection has problems such as low efficiency, high cost, and inconsistent standards. The project "Automatic Recognition of Snack Packaging Time Information Based on YOLOv11" aims to solve this problem by focusing on the unique challenges of recognizing snack packaging time information in bags and boxes. Based on advanced YOLOv11 deep learning algorithms, a high-precision and highly robust end-to-end intelligent recognition system is constructed. At the technical level, C3k2 module is used to replace traditional C2f module, and attention mechanism (C2PSA module) is introduced. The detection and recognition module is integrated into a single network head to build an end-to-end pipeline. Hardware technology is integrated to achieve fast and accurate extraction of date information in complex packaging scenarios. In terms of data collection, in response to the problem that the production date and shelf life of most product packaging cannot be in one photo, tools such as Photoshop are used to clip the two in the same image. In addition, the dataset covers various materials and real packaging information with interference such as wrinkles and reflections. Through data augmentation, the original dataset is scaled and rotated to ensure its diversity. The results show that the model can quickly and accurately identify the production date and shelf-life information on food packaging in real-time within 0.5 seconds, and recognize the text of the box selection information through text recognition technology, and finally calculate whether the food has expired. This project implements YOLOv11 architecture improvement and end-to-end lightweight design at the technical level, and has multi scenario adaptability at the application level. Firstly, it can help consumers solve the problem of difficult identification of food packaging information; In addition, this technology is expected to be extended and applied in retail logistics and warehousing scenarios, as well as in the fields of pharmaceuticals and cosmetics.