Data Processing Scheme for Laser Spot Thermography Applied for Nondestructive Testing of Composite Laminates

This paper proposes a data processing scheme for laser spot thermography (LST) applied for nondestructive testing (NDT) of composite laminates. The LST involves recording multiple thermographic sequences, resulting in large amounts of data that have to be processed cumulatively to evaluate the diagnostic information. This paper demonstrates a new data processing scheme based on parameterization and machine learning. The approach allows to overcome some of the major difficulties in LST signal processing and deliver valuable diagnostic information. The effectiveness of the proposed approach is demonstrated on an experimental dataset acquired for a laminated composite sample with multiple simulated delaminations. The paper discusses the theoretical aspects of the proposed signal processing and inference algorithms as well as the experimental arrangements necessary to collect the input data.