Research on the Application of an Offline Equipment Data Real-time Acqui-sition System Based on Visual Recognition and Large Language Model

To address the issue of outdated manufacturing equipment lacking internet connectivity and failing to meet real-time data collection requirements for on-site device management,quality control,and digital systems,an offline equipment data real-time collection system based on visual recognition and large language model is proposed and designed.The system adopts a five-layer architecture,including hardware collection,OCR recognition,intelligent parsing,data access,and business application layers.Field validation demonstrates its strong stability and accuracy,enabling real-time high-frequency offline device operation data collection and anomaly closed-loop handling at low cost without modifying the original equipment's electrical structure.It provides a reusable engineering paradigm for equipment reuse and intelligent upgrading in the manufacturing industry.