Terahertz ray imaging is one of the most capable techniques to inspect the space shuttle external tank foam insulation. This technique, however, is limited in its current inspection protocol using indirect substrate reflections. An alternate signal processing approach, working directly on the flaw responses, was lately demonstrated to be able to overcome some of these limitations. In this paper, we report recent progresses made by utilizing this alternate signal processing procedure in additional samples to detect flaws that were missed by the current protocol. We also present a new detection approach using the probabilistic neural network in the context of Bayesian classification. Preliminary results showed that the new Bayesian classification approach can achieve even greater improvement over the alternate signal processing approach.
[1]
Eric I. Madaras,et al.
Modeling and Processing of Terahertz Imaging in Space Shuttle External Tank Foam Inspection
,
2006
.
[2]
D. F. Specht,et al.
Probabilistic neural networks for classification, mapping, or associative memory
,
1988,
IEEE 1988 International Conference on Neural Networks.
[3]
David Zimdars,et al.
High Resolution Co-Linear Stand-off Terahertz Reflection Imaging
,
2005
.
[4]
William P. Winfree,et al.
Technology and Applications of Terahertz Imaging Non‐Destructive Examination: Inspection of Space Shuttle Sprayed On Foam Insulation
,
2005
.