Implementation of advanced signal processing techniques on Line-Scan Thermography data

In the last few years, composite materials have found an important niche of application in several industries, mainly because of their improved mechanical properties (higher stiffness, strength and resistance to fatigue). In this context, sandwich-composites, a special class of composite materials - are commonly used in the aerospace industry to manufacture lighter components. The increasing use of this type of materials in the aerospace sector has opened the necessity of inspection methods to evaluate its physical integrity and quality. Line Scan Thermography (LST) is one of the emerging technologies aimed to detect and evaluate subsurface defects present in the sandwiches composite structures. As a non-destructive testing and evaluation (NDT&E) technique, LST is a dynamic technique suited to inspect large and complex aerospace components. However, its performance to detect deeper and smaller defects is negatively affected due to the different sources of noise present in the collected thermal images. In this paper is studied the application of advanced signal processing techniques on LST data obtained from the inspection of a large composite component, which contains different types of internal defects located at a variety of depths. To evaluate the ability of each technique to reduce the noise, the signal-to-noise ratio (SNR) at the maximum signal contrast of each defect has been computed for further analysis.