AN ARTIFICIAL NEURAL-NETWORK INTERPRETER FOR TRANSIENT THERMOGRAPHY IMAGE DATA

Authors
SAINTEY MB ALMOND DP
Citation
Mb. Saintey et Dp. Almond, AN ARTIFICIAL NEURAL-NETWORK INTERPRETER FOR TRANSIENT THERMOGRAPHY IMAGE DATA, NDT & E international, 30(5), 1997, pp. 291-295
Citations number
9
Categorie Soggetti
Materials Science, Characterization & Testing
Journal title
NDT & E international → ACNP
ISSN journal
09638695
Volume
30
Issue
5
Year of publication
1997
Pages
291 - 295
Database
ISI
SICI code
0963-8695(1997)30:5<291:AANIFT>2.0.ZU;2-A
Abstract
It is shown that an artificial neural network can be used as an expert system to obtain details of defect size and depth from transient ther mography data. The neural network was trained on a comprehensive range of finite difference modelling results and is shown to provide a simp le means of interpreting practical experimental measurements. An asses sment is given of the sensitivity of the method to likely uncertaintie s in experimental measurement. (C) 1997 Elsevier Science Ltd.