Wavelet transform based post processing approach for pulse compression favourable frequency modulated thermal wave imaging for non-destructive testing and evaluation

Abstract

InfraRed Thermography (IRT) for Non-Destructive Testing and Evaluation (NDT&E) is used across various industries to inspect components and structures to detect and identify the hidden defects in them. Since Thermal NDT&E (TNDT&E) captures the surface conditions remotely within a reasonable time, several inspection and predictive maintenance programs have employed challenging traditional computer vision algorithms to detect surface and subsurface defects in various materials. In this work, a Glass Fibre Reinforced Polymer (GFRP) sample having artificially simulated flat-bottom holes at various depths was examined with the Frequency Modulated Thermal Wave Imaging (FMTWI) technique. The thermal data acquired using FMTWI for the sample contains various environmental and experimentally generated noises. Principal Component Thermography (PCT) using Singular Value Decomposition (SVD) has been carried out to qualitatively enhance the spatial thermal contrast that simultaneously provides temporal thermal profiles with reduced dynamic range. Pulse Compression on the obtained PCT sequence using the cross-correlation approach improves the range resolution of the subsurface defects. Wavelet transforms like Haar, Mexican Hat, and Morlet are applied to the frames within the half-width region of the main lobe in pulse-compressed signal to study their effect on defect detection capabilities and visualize the Region of Interest (RoI), specifically the circular defect region within the material. The novel thermal pattern enhancement method extends IRT-based Automated Defect Recognition (ADR) for more precise defect-detection and localization particularly for GFRP structures, which are mostly used in aerospace and shipbuilding industries.