当前位置: X-MOL 学术J. Nondestruct. Eval. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Augmented Ultrasonic Data for Machine Learning
Journal of Nondestructive Evaluation ( IF 2.8 ) Pub Date : 2021-01-02 , DOI: 10.1007/s10921-020-00739-5
Iikka Virkkunen , Tuomas Koskinen , Oskari Jessen-Juhler , Jari Rinta-aho

Flaw detection in non-destructive testing, especially for complex signals like ultrasonic data, has thus far relied heavily on the expertise and judgement of trained human inspectors. While automated systems have been used for a long time, these have mostly been limited to using simple decision automation, such as signal amplitude threshold. The recent advances in various machine learning algorithms have solved many similarly difficult classification problems, that have previously been considered intractable. For non-destructive testing, encouraging results have already been reported in the open literature, but the use of machine learning is still very limited in NDT applications in the field. Key issue hindering their use, is the limited availability of representative flawed data-sets to be used for training. In the present paper, we develop modern, deep convolutional network to detect flaws from phased-array ultrasonic data. We make extensive use of data augmentation to enhance the initially limited raw data and to aid learning. The data augmentation utilizes virtual flaws—a technique, that has successfully been used in training human inspectors and is soon to be used in nuclear inspection qualification. The results from the machine learning classifier are compared to human performance. We show, that using sophisticated data augmentation, modern deep learning networks can be trained to achieve human-level performance.

中文翻译:

用于机器学习的增强超声数据

无损检测中的缺陷检测,尤其是超声波数据等复杂信号,迄今为止在很大程度上依赖于训练有素的人类检查员的专业知识和判断力。虽然自动化系统已经使用了很长时间,但它们大多仅限于使用简单的决策自动化,例如信号幅度阈值。各种机器学习算法的最新进展已经解决了许多以前被认为难以处理的类似困难的分类问题。对于无损检测,公开文献中已经报道了令人鼓舞的结果,但机器学习在该领域的无损检测应用中的使用仍然非常有限。阻碍其使用的关键问题是用于训练的代表性缺陷数据集的可用性有限。在本文中,我们开发了现代的深度卷积网络来检测相控阵超声数据中的缺陷。我们广泛使用数据增强来增强最初有限的原始数据并帮助学习。数据增强利用虚拟缺陷——一种已成功用于培训人类检查员的技术,很快将用于核检查资格。将机器学习分类器的结果与人类表现进行比较。我们表明,使用复杂的数据增强,可以训练现代深度学习网络以实现人类水平的表现。这已成功用于培训人类检查员,并将很快用于核检查资格。将机器学习分类器的结果与人类表现进行比较。我们表明,使用复杂的数据增强,可以训练现代深度学习网络以实现人类水平的表现。这已成功用于培训人类检查员,并将很快用于核检查资格。将机器学习分类器的结果与人类表现进行比较。我们表明,使用复杂的数据增强,可以训练现代深度学习网络以实现人类水平的表现。
更新日期:2021-01-02
down
wechat
bug