An Augmented Reality and AI-Driven Framework for Real-Time Analog Gauge Digitization in Industrial Environments

Manual analog gauge reading is error-prone, inefficient, and incompatible with modern digital monitoring systems, limiting its effectiveness in real-time industrial automation. To address these limitations, this paper proposes a real-time AI and Augmented Reality (AR)-driven framework for automated analog gauge digitization. The system integrates YOLOv8s for high-speed gauge localization, a Swin Transformer-enhanced UNet for precise pointer segmentation, and a Kalman filter-based module for stabilized angle-to-value conversion. A Unity-based AR interface overlays digitized readings onto physical gauges in real time, enhancing operator awareness and interaction. The framework achieves a detection accuracy of 95.6% mAP@0.5 and maintains a mean absolute error (MAE) of 1.87 units on real-world datasets. It is optimized for deployment on edge devices such as NVIDIA Jetson Nano and Google Pixel 6, operating with an end-to-end latency of less than 150 milliseconds. Usability testing reports a high System Usability Score (SUS) of 86.2/100 and accurate AR alignment with an average deviation of 6.1 pixels and 2.3 degrees. Overall, the proposed framework offers a practical, accurate, and scalable solution for intelligent industrial monitoring through seamless AI-AR integration.