Research On Defect Visualization Method Based On Eddy Current Thermography

In the safety production,it is of great significance to inspect the structural damage and ensure the safe service of the structure,which can avoid accidents to the maximum extent.Eddy current thermography is a non-contact measurement of non-destructive testing technology that can effectively detect equipment structure.In this thesis,the moving mode of eddy current thermography is studied around the defect temperature signal reconstruction,image denoising and contour reconstruction.The main contents are as follows:(1)For the moving mode of eddy current thermography,a signal reconstruction method based on area scanning is proposed,the position of the optimal scanning area is determined,then the area division criteria and operation methods are analyzed to realize the defect temperature signal reconstruction in the moving detection mode.(2)For the reconstructed image obtained by the area scanning method,a noise reduction method based on a combination of lifting wavelet improved threshold function and Wiener filtering is used.An adaptive threshold and an improved threshold function are selected to enhance the high-frequency part obtained by the lifting wavelet decomposition.And the Wiener filtering method is used to denoise the low-frequency part.Compare the results with single wavelet threshold denoising method and Wiener filtering method to verify the effectiveness of the method.(3)Based on the analysis of the relationship between defect depth,diameter and temperature,a nonlinear mapping of defect temperature contour and two-dimensional geometric contour is established by using the method of least squares support vector machine based on particle swarm optimization.The two-dimensional contours of defects are reconstructed,which achieve the visualization of defects.