Sparse Principal Component Thermography for Subsurface Defect Detection in Composite Products
| Autor: | Jin-Yi Wu, Yuan Yao, Stefano Sfarra |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Computer science
Active thermography data analysis nondestructive testing (NDT) sparse principal component analysis (SPCA) thermographic analysis data analysis Carbon fibers 02 engineering and technology sparse principal component analysis (SPCA) thermographic analysis Nondestructive testing 0202 electrical engineering electronic engineering information engineering Electrical and Electronic Engineering Active thermography business.industry Noise (signal processing) 020208 electrical & electronic engineering nondestructive testing (NDT) Pattern recognition Fibre-reinforced plastic 021001 nanoscience & nanotechnology Computer Science Applications Characterization (materials science) Control and Systems Engineering visual_art Principal component analysis Thermography visual_art.visual_art_medium Artificial intelligence 0210 nano-technology business Information Systems |
| Zdroj: | IEEE Transactions on Industrial Informatics. 14:5594-5600 |
| ISSN: | 1941-0050 1551-3203 |
| DOI: | 10.1109/tii.2018.2817520 |
| Popis: | Active thermography is an efficient and powerful technique for nondestructive testing of products made of composite materials, which enables rapid inspection of large areas, presents results as easily interpreted high-resolution images, and is easy to operate. In recent years, a number of thermographic data analysis methods were developed to enhance the visibility of subsurface defects, among which principal component thermography (PCT) is recommended because of its capability to enhance the contrast between defective and defect-free areas, compress data, and reduce noise. In this study, a sparse principal component thermography (SPCT) method is proposed, which inherits the advantages of PCT and allows more flexibility by introducing a penalization term. Compared to PCT, SPCT provides more interpretable analysis results owing to its structure sparsity. The feasibility and effectiveness of the proposed method are illustrated by the experimental results of the subsurface defect characterization in a carbon fiber reinforced plastic specimen. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|