Backpropagation learning algorithm for nondestructive testing by thermal imager [aerospace materials]

Author

Kojima, Fumio ; Kawaguchi, Hiroshi

Author_Institution

Dept. of Mech. Eng., Osaka Inst. of Technol., Japan

Volume

1

fYear

1993

fDate

25-29 Oct. 1993

Firstpage

955

Abstract

An artificial neural network is applied to nondestructive inspections using thermal imager. The use of an artificial neural network is presented for classifying test data as corresponding to bonded and disbonded regions in sample materials. The backpropagation learning for a multi-layer feedforward neural network is applied to this classification. The trust region method is adopted to the backpropagation learning problem. Results of numerical tests are summarized.

Keywords

aerospace computing; aircraft testing; backpropagation; feedforward neural nets; flaw detection; image classification; infrared imaging; multilayer perceptrons; nondestructive testing; aerospace materials; artificial neural network; backpropagation learning algorithm; bonded regions; disbonded regions; multi-layer feedforward neural network; nondestructive testing; test data classification; thermal imager; trust region method; Aerospace materials; Aerospace testing; Artificial neural networks; Backpropagation algorithms; Bonding; Inspection; Materials testing; Multi-layer neural network; Neural networks; Nondestructive testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 1993. IJCNN '93-Nagoya. Proceedings of 1993 International Joint Conference on

Print_ISBN

0-7803-1421-2

Type

conf

DOI

10.1109/IJCNN.1993.714069

Filename

714069