Coating Thickness Measurements and Defect Characterization in Non-Metallic Composite Materials by Using Thermography

Abstract

Thermography is a non-destructive testing method (NDT), which is widely used to guarantee the quality of non-metallic materials, such as carbon fiber composite, anti-reflection (AR) film, and coatings. As other NDT methods do, thermography determines a defective area based on the signal difference between suspected defective areas and defective-free areas. Two unavoidable effects are decreasing the credibility of thermography detection: one is uneven heating, and the other is lateral diffusion of heat. To solve this problem, researchers have developed various reconstruction methods. Restoring methods are known to have the capacity to reduce the effect of heat-flux lateral diffusion by de-convoluting a point spread function either along a temporal profile or a spatial profile to process captured thermal images. These methods either require pre-knowledge with depth or are not effective in detecting deep defects. Here we propose a spatial-temporal profile-based reconstruction method to reduce the effect of uneven heating and lateral diffusion. The method evaluates the heat flux deposited onto tested samples based on surface temperature gathered under ideal conditions. Then the proposed method is tested in three real applications – in defect detection on semi-transparent materials, on semi-infinite defects (coatings) and anisotropic materials. The method is evaluated against existing methods. Results suggest that the proposed method is effective and computationally efficiently over all the reconstruction methods reviewed. It reduces the effect of uneven heating by providing a good approximation to the input heat flux at the ending image of the sequence.