Comparative analysis on thermal non-destructive testing imagery applying Candid Covariance-Free Incremental Principal Component Thermography (CCIPCT)

Bardia Yousefi, Stefano Sfarra, Clemente Ibarra Castanedo, Xavier P.V. Maldague

Research output: Contribution to journalArticle

38 Citations (Scopus)


Thermal and infrared imagery creates considerable developments in Non-Destructive Testing (NDT) area. Here, a thermography method for NDT specimens inspection is addressed by applying a technique for computation of eigen-decomposition which refers as Candid Covariance-Free Incremental Principal Component Thermography (CCIPCT). The proposed approach uses a shorter computational alternative to estimate covariance matrix and Singular Value Decomposition (SVD) to obtain the result of Principal Component Thermography (PCT) and ultimately segments the defects in the specimens applying color based K-medoids clustering approach. The problem of computational expenses for high-dimensional thermal image acquisition is also investigated. Three types of specimens (CFRP, Plexiglas and Aluminium) have been used for comparative benchmarking. The results conclusively indicate the promising performance and demonstrate a confirmation for the outlined properties.

Original languageEnglish
Pages (from-to)163-169
Number of pages7
JournalInfrared Physics and Technology
Publication statusPublished - 1 Sep 2017



  • Candid covariance-free incremental principal component thermography
  • K-Medoids clustering
  • Non-Destructive Testing (NDT)
  • Principal component thermography
  • Thermal image analysis

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Atomic and Molecular Physics, and Optics
  • Condensed Matter Physics

Cite this