Automated defect detection based on time and frequency domain analysis using Eddy Current Pulsed Thermography in EB and TIG welding inspections

R. Hidalgo-Gato, E. Gorostegui-Colinas, P. López de Uralde, A. Muniategui, B. Urtasun, I. Silanes
2020 Proceedings of the 2020 International Conference on Quantitative InfraRed Thermography   unpublished
On this work an active thermography has been applied to Inconel 718 EBW and TIG samples, continuing previous research on this field [1, 2] . In particular, the objective of this work is to improve and automate the processing part of the automatic defect detection procedure to enhance the results of the detection algorithm, which is the one applied in the last step. The proposed method is based on a time and frequency domain analysis (TFDA), in which, the resulting thermograms contain the most
more » ... levant thermal information regarding to the defect thermal evolution. To assess the suitability on each resulting thermogram (including the fused images), an accuracy evaluation based on Tanimoto criterion has been employed. This evaluation leads to an optimate processing procedure that ensures a detection with less false positives or negatives.
doi:10.21611/qirt.2020.023 fatcat:lhmo7fotsncalnubwpsyog7wry