Invitation to the lecture within the cycle „Polish-German Scientific Encounters”
Prof. Marcin Hinz from the Faculty of Mechanical, Automotive and Aeronautical Engineering, Munich University of Applied Sciences will give a lecture in English on AI in Engineering: Advanced Methods for Assessing Surface Quality in Manufacturing Processes.
The event will be facilitated by:
- Prof. Ewa Niewiadomska-Szynkiewicz, Associate Dean for Science, Faculty of Electronics and Information Technology, Warsaw University of Technology
- Prof. Robert Nowak, Deputy Director for Research, Institute of Computer Science, Faculty of Electronics and Information Technology, Warsaw University of Technology
Date: March 27, 2025 (Thursday), 3:00 PM - 4:30 PM
Location: Faculty of Electronics and Information Technology, Central Auditorium (1st floor).
Prior registration is required with an email address in the WUT domain.
Registration form: HERE.
Abstract:
The integration of Artificial Intelligence (AI) in engineering quality assessment is transforming manufacturing and production processes by enhancing efficiency, accuracy, and scalability. Traditional quality control methods, which rely on manual inspections and tactile measurements, are often time-consuming, subjective, and prone to human error. AI-driven solutions, leveraging machine learning and computer vision, enable automated, precise, and real-time quality assessment, reducing costs and minimizing production downtime.
This study focuses on the application of machine learning for surface quality analysis, particularly in evaluating surface roughness in high-precision manufacturing. Conventional surface measurement techniques, such as tactile roughness testers and optical 3D microscopes, are limited by their time-intensive and costly nature. To address these challenges, an AI-based framework is introduced, integrating image-based analysis, computer vision, and machine learning techniques for efficient surface inspection.
The experimental setup involves capturing high-resolution images of product surfaces under controlled lighting conditions, followed by image pre-processing and feature extraction. Various machine learning models—including supervised, unsupervised, and semi-supervised approaches—are assessed for their effectiveness in surface classification and regression tasks. Supervised learning methods such as Random Forest, Support Vector Machines, Neural Networks, and Long Short-Term Memory (LSTM) models are employed to enhance classification accuracy, while convolutional neural networks (CNNs) demonstrate strong performance in identifying surface characteristics. As AI continues to advance, its role in quality assessment is expected to further improve manufacturing efficiency, adaptability, and reliability. This research contributes to the development of AI-driven surface inspection systems, providing a scalable and precise alternative to traditional measurement techniques in engineering and production environments.