Welding Defect Detection Using CNNs: Improving Non-Destructive Testing – A Case Study of an Algerian Company

DOI

10.2991/978-94-6463-805-9_24How to use a DOI?

Keywords

welding defects YOLO v8; Convolutional Neural network (CNN) Non-Destructuve Testing; Pretrained Models; Real Time Detection System

Abstract

Welding plays a pivotal role across various industries as an assembly process. During its execution, certain defects and imperfections are likely to emerge. To rectify these imperfections without compromising the integrity of the components, techniques such as non-destructive testing through radiology are employed. However, this labor-intensive and costly manual analysis process is time-consuming, leading to increased expenses.

Industry 4.0 refers to the gradual transformation of traditional manufacturing processes through the integration of digital technologies, automation, data exchange, and advanced analytics, creating new opportunities.

Our study culminates in the development of A real-time welding defect detection system based on Convolutional Neural Networks (CNNs). This system leverages object detection technology and employs pre-trained models such as YOLOv7, YOLOv8, Faster R-CNN, and RetinaNet to automate the non-destructive testing process. The image dataset was curated from two data sources, the first of which was obtained through collaboration with the Algerian company Eurl TESTIAL, and the second one is a public dataset called GDXray. The achieved results exhibit a precision rate of up to 87.50% and a Mean Average Precision (mAP) of 84.60% using YOLOv8.

This system offers significant advantages in terms of efficiency, accuracy, and resource savings within the realm of industrial inspection. By seamlessly integrating Industry 4.0 principles and object detection, our work contributes to the enhancement of quality control processes within welding and assembly operations.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Fatima Kabli
AU  - Amal Boumadjout
AU  - Houssam Abdelhakim Turki
AU  - Djelloul Ferrah
AU  - Mohamed Mokhtari
PY  - 2025
DA  - 2025/08/05
TI  - Welding Defect Detection Using CNNs: Improving Non-Destructive Testing – A Case Study of an Algerian Company
BT  - Proceedings of the First International Conference on Artificial Intelligence, Smart Technologies and Communications (AISTC 2025)
PB  - Atlantis Press
SP  - 211
EP  - 219
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6463-805-9_24
DO  - 10.2991/978-94-6463-805-9_24
ID  - Kabli2025
ER  -