Quantification of depth and morphology of internal corrosion defects by stepped eddy current thermography skewness under weak excitation conditions

Abstract

In the nondestructive testing (NDT) of eddy current thermography of oil and gas tanks, the low and non-uniform temperature rise due to weak excitation conditions makes the quantitative estimation of the depth and area of defects still challenging. This work proposes a mathematical-physical model of skewness to characterize defect depth quantitatively. The skewness image is estimated in Gaussian scale space using the background difference method combined with the chunked threshold segmentation method to characterize the defect morphology. The experimental results show that the skewness is affected by the combination of defect depth and area. As the defect area decreases, the relationship between defect depth and skewness gradually degrades from nonlinear to linear, while the rate of change of skewness becomes insignificant. The skewness analysis model predicted the maximum MRE of 5.15%, the minimum R2 of 0.94 for defect depth, and the MRE of 8.39% for defect area. It shows that the skewness can effectively characterize the depth and morphology of internal corrosion defects.