VISUAL DATA ACQUISITION USING YOLO AND OCR FOR SPUTTRING PROCESS MONITORING

This paper presents a real-time data visualization and fault detection model for sputtering process monitoring, focusing on parameters such as deposition rate, film thickness, material types (Ti, Ag, Ni), and process status. The objective is to model multi-level anomaly outliers for detecting potential OCR errors and sputtering process deviations, develop a real-time embedded vision system for automated data acquisition from the sputtering equipment's display, and implement a complete monitoring application with real-time visualization and post-process analysis for industrial deployment. Data acquisition is carried out using a high-resolution camera, where YOLO achieves 99.5% [email protected] in supervised detection of visual indicators, and PaddleOCR attains 99.57% accuracy in extracting numerical parameters. Preprocessing incorporates a median filter to suppress noise, while DBSCAN identifies sudden OCR fluctuations and linear regression models parameter trends. The postprocessed data are stored in structured CSV files. By integrating supervised and unsupervised learning with data science techniques, the proposed system enables reliable monitoring, early anomaly detection, and predictive maintenance in industrial sputtering operations.