Title :
A neural network approach to ultrasonic tomography
Author :
Hutchins, D.A. ; Mottram, J.T. ; Hines, E.L. ; Corcoran, P. ; Anthony, D.M.
Author_Institution :
Dept. of Eng., Warwick Univ., Coventry, UK
Abstract :
An artificial neural network (ANN) is used to form a simple tomographic image from ultrasonic data. Data are taken using a pinducer array on one face of a fiber-reinforced polymer composite plate. The system is shown to detect artificial defects in the form of holes through the composite sample
Keywords :
acoustic signal processing; backpropagation; carbon fibre reinforced plastics; computerised tomography; crack detection; delamination; feedforward neural nets; image recognition; physics computing; ultrasonic materials testing; C; NDE; artificial defects; artificial neural network; backpropagation; delamination cracks detection; fiber-reinforced polymer composite plate; holes; multilayer preceptron; ultrasonic tomography; Artificial neural networks; Data analysis; Delamination; Image reconstruction; Instruments; Laminates; Neural networks; Polymers; Tomography; Ultrasonic imaging;
Conference_Titel :
Ultrasonics Symposium, 1992. Proceedings., IEEE 1992
Conference_Location :
Tucson, AZ
Print_ISBN :
0-7803-0562-0
DOI :
10.1109/ULTSYM.1992.275981
