COMPARISON OF SELECTED WELD DEFECT EXTRACTION METHODS

Journal Article · Thu Feb 28 00:00:00 EST 2008 · AIP Conference Proceedings

DOI:https://doi.org/10.1063/1.2902546· OSTI ID:21210226

Sikora, R; Baniukiewicz, P; Chady, T; Rucinski, W; Swiadek, K; Caryk, M; Lopato, P ^[1]

---

1. Szczecin University of Technology, Department of Electrical Engineering, 70-313 Szczecin (Poland)

This paper presents three different methods of welding defects detection from radiographs. First two methods are dedicated for extraction of flaws directly from radiograms. Fuzzy logic system considers whether the pixel belongs to crack (or background) using defects probability maps together with simple fuzzy rules. Neural network method uses knowledge gathered from radiographs with known defects. Indirect method extracts defects by subtracting background from the radiograph and next using various local thresholding methods.

• OSTI ID: 21210226
• Journal Information: AIP Conference Proceedings, Vol. 975, Issue 1; Conference: 34. annual review of progress in quantitative nondestructive evaluation, Golden, CO (United States), 22-27 Jul 2007; Other Information: DOI: 10.1063/1.2902546; (c) 2008 American Institute of Physics; Country of input: International Atomic Energy Agency (IAEA); ISSN 0094-243X
• Country of Publication: United States
• Language: English

Related Subjects

36 MATERIALS SCIENCE
APPROXIMATIONS
CRACKS
DEFECTS
FUZZY LOGIC
GAMMA RADIOGRAPHY
IMAGE PROCESSING
IMAGES
NEURAL NETWORKS
PROBABILITY
WELDED JOINTS
WELDING
X-RAY RADIOGRAPHY