DEVELOPMENT OF A COMPUTER VISION-BASED SYSTEM FOR AUTOMATIC IDENTIFICATION AND REGISTRATION OF PART SERIAL NUMBERS IN INDUSTRIAL ENTERPRISES

Abstract

Abstract. The rapid growth of Industry 4.0 technologies and the increasing demand for digital transformation have created a need for automated identification and traceability systems in industrial enterprises. Serial numbers assigned to mechanical parts and products play a crucial role in production control, quality assurance, inventory management, and lifecycle tracking. However, traditional manual registration methods are time-consuming, susceptible to human error, and inefficient when dealing with large volumes of products. To address these challenges, this study proposes a computer vision-based system for the automatic identification and registration of part serial numbers in industrial environments. The developed system integrates image acquisition devices, image preprocessing techniques, optical character recognition (OCR), and a database management module to ensure accurate extraction and storage of serial number information. The methodology includes image enhancement, noise reduction, segmentation, and text recognition using deep learning-based OCR algorithms. Experimental evaluation demonstrates that the proposed approach provides high recognition accuracy, reduces processing time, and minimizes human intervention compared with conventional methods. The implementation of such a system contributes to improved operational efficiency, enhanced product traceability, and increased reliability of manufacturing processes, making it suitable for modern smart factories and industrial automation systems.