Ensemble Joint Sparse Low-Rank Matrix Decomposition for Thermography Diagnosis System

Abstract

Composite is widely used in the aircraft industry and it is essential for manufacturers to monitor its health and quality. The most commonly found defects of composite are debonds and delamination. Different inner defects with complex irregular shape are difficult to diagnosed by using conventional thermal imaging methods. In this article, an ensemble joint sparse low-rank matrix decomposition algorithm is proposed by applying the optical pulse thermography (OPT) diagnosis system. The proposed algorithm jointly models the low-rank and sparse pattern by using concatenated feature space. In particular, the weak defects information can be separated from strong noise and the resolution contrast of the defects has significantly been improved. Ensemble iterative sparse modeling are conducted to further enhance the weak information as well as reducing the computational cost. In order to show the robustness and efficacy of the model, experiments are conducted to detect the inner debond on multiple carbon fiber reinforced polymer composites. A comparative analysis is presented with general OPT algorithms. Notwithstanding above, the proposed model has been evaluated on synthetic data and compared with other low-rank and sparse matrix decomposition algorithms.