An efficient data processing approach for frequency modulated thermal wave imaging for inspection of steel material

Abstract

Non-destructive testing & evaluation (NDT&E) plays a vital role in industrial quality control. Among various NDT&E modalities, active thermal NDT&E gained its importance due to its inherent merits such as remote, whole-field, fast and quantitative inspection capabilities. Of various thermal NDT&E schemes, recently proposed pulse compression favorable frequency modulated thermal wave imaging (FMTWI) became popular due to its enhanced defect detection sensitivity along with improved test resolution. This paper presents noise rejection capabilities of FMTWI with principal component analysis (PCA) based post-processing schemes. PCA based postprocessing helps in efficient interpretation of the thermographic data by removing artefacts and producing few significant images depicting sub-surface defects in the test specimen. The results obtained by PCA can be made more interpretable by using sparser version of PCA (SPCA). In this paper, SPCA based thermographic data processing technique is proposed in which SPCA has been considered in two different ways. Firstly it has been implemented to induce sparsity in empirical orthogonal functions (EOFs) which improves spatial contrast over the defective regions. Secondly, it has been used to modify principal components (PCs) (time series components) to obtain resultant images by projecting thermographic data on modified PCs which manages to enhance the signal to noise ratio (SNR). The sub-surface defect detection capabilities of the proposed methods are studied by a matched filter based pre-processing scheme which reduces the computational cost and memory usage also. The performance of the proposed methods has been evaluated on the experimental investigation of the mild steel specimen having flat bottom holes as defects.