Automatic quality assurance of the smart factory’s products with use of computer vision methods

Abstract

The advent of Industry 4.0 has revolutionized manufacturing by integrating automation and data-driven technologies into smart factories. Quality assurance (QA) remains a critical challenge, where manual inspection methods are often slow, error-prone, and inefficient. This thesis explores the application of computer vision to automate QA processes in a smart factory environment, leveraging Basler's cameras for high-precision image acquisition and OpenCV for real-time defect detection. The proposed system is designed to interface with Programmable Logic Controllers (PLCs) via TIA Portal, enabling seamless integration with existing industrial automation infrastructure. A proof-of-concept (POC) system is developed to detect defects such as surface scratches, misalignments, or missing components on production lines. The workflow includes image capture via Basler cameras, preprocessing and defect classification using OpenCV-based algorithms, and triggering corrective actions through PLCs. The system’s performance is evaluated based on accuracy, latency, and scalability, demonstrating its potential to reduce human intervention while improving QA efficiency.