Autoencoder-Based Anomaly Detection in Industrial X-ray Images
2021 (English)In: Proceedings of 2021 48th Annual Review of Progress in Quantitative Nondestructive Evaluation. Virtual, Online. July 28–30, 2021., ASME Press, 2021, p. 28-30, article id V001T07A001Conference paper, Oral presentation with published abstract (Refereed)
Abstract [en]
Within many quality-critical industries, e.g. the aerospace industry, industrial X-ray inspection is an essential as well as a resource intense part of quality control. Within such industries the X-ray image interpretation is typically still done by humans, therefore, increasing the interpretation automatization would be of great value. We claim, that safe automatic interpretation of industrial X-ray images, requires a robust confidence estimation with respect to out-of-distribution (OOD) data. In this work we have explored if such a confidence estimation can be achieved by comparing input images with a model of the accepted images. For the image model we derived an autoencoder which we trained unsupervised on a public dataset with X-ray images of metal fusion-welds. We achieved a true positive rate at 80–90% at a 4% false positive rate, as well as correctly detected an OOD data example as an anomaly.
Place, publisher, year, edition, pages
ASME Press, 2021. p. 28-30, article id V001T07A001
Keywords [en]
X-rays, Aerospace industry, Inspection, Metals, Quality control, Welded joints
National Category
Manufacturing, Surface and Joining Technology
Research subject
Production Technology
Identifiers
URN: urn:nbn:se:hv:diva-18231DOI: 10.1115/qnde2021-74428Scopus ID: 2-s2.0-85123054602ISBN: 978-0-7918-8552-9 (print)OAI: oai:DiVA.org:hv-18231DiVA, id: diva2:1643536
Conference
48th Annual Review of Progress in Quantitative Nondestructive Evaluation, QNDE 2021. Virtual, Online. July 28–30, 2021.
Funder
ÅForsk (Ångpanneföreningen's Foundation for Research and Development)2022-03-102022-03-102022-03-29Bibliographically approved