Defect depth estimation in passive thermography using neural network paradigm

Defect depth estimation from passive thermography data based on neural network paradigm is proposed. Three parameters have been found to be related with depth of the defect. Therefore, these parameters: the maximum temperature over the defective area (T-max), the temperature on the non-defective or sound area (T-so), and the average temperature (T-avg) of the inspected area have been used as input parameters to train multilayer perceptron neural networks. For verification of the proposed scheme, NN has been tested with trained and untrained data. The correct depth estimation is 100% for trained data and more than 98% for untrained data. The result shows a great potential of the proposed method for defect depth estimation by means of passive thermography.

[1]  Herbert Kaplan,et al.  Practical Applications of Infrared Thermal Sensing and Imaging Equipment , 1993 .

[2]  Cataldo Guaragnella,et al.  Defect detection in aircraft composites by using a neural approach in the analysis of thermographic images , 2005 .

[3]  Martin T. Hagan,et al.  Neural network design , 1995 .

[4]  H. Qi,et al.  Clinical study on using thermal texture maps in SARS diagnosis , 2004, The 26th Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[5]  Xavier Maldague,et al.  Theory and Practice of Infrared Technology for Nondestructive Testing , 2001 .

[6]  Xavier Maldague,et al.  Neural network based defect detection and depth estimation in TNDE , 2002 .

[7]  X. Maldague.  Nondestructive Evaluation of Materials by Infrared Thermography , 1993 .

[8]  Xavier Maldague,et al.  Defect characterization in pulsed thermography: a statistical method compared with Kohonen and Perceptron neural networks , 2000 .

[9]  Hairong Qi,et al.  Early detection of breast cancer using thermal texture maps , 2002, Proceedings IEEE International Symposium on Biomedical Imaging.

[10]  Syed Abdul Rahman,et al.  Modelling of defects within high temperature wall by means of infrared thermography , 2007 .

[11]  D. P. Almond,et al.  An artificial neural network interpreter for transient thermography image data , 1997 .

[12]  Xavier Maldague,et al.  A study of defect depth using neural networks in pulsed phase thermography: modelling, noise, experiments , 1998 .

[13]  Ermanno G. Grinzato,et al.  Application of neural network computing to thermal non-destructive evaluation , 1997, Neural Computing & Applications.