A proposPersonnesRecherchePublicationsEvenementsProfil
A propos
Séminaires REPARTI

Les Séminaires REPARTI à l'Université Laval ont lieu le vendredi à 11h30.
Veuillez consulter le programme pour plus de détails.

Projet de maîtrise, de doctorat ou stage postdoctoral en apprentissage automatique au sein de l'équipe du Prof. Christian Gagné : veuillez consulter l'annonce suivante pour tous les détails : http://vision.gel.ulaval.ca/~cgagne/postes2017.html







Mar 14 2008 11:30AM

Matthieu Klein

Thermographic signal processing through correlation operators in pulsed thermography


In active non-destructive testing by IR thermography, the sample to be tested is usually excited with an external source of heat. The source is either modulated in the case of lock-in thermography, or instantaneous in the case of pulsed thermography (PT). In both cases the transient state of heat diffusion is recorded as a sequence of N infrared images for a duration of time t. The biggest defects are obvious and immediately noticeable from single raw IR images. To detect smaller defects within the sample, efficient enhancement methods of the signature of the heat diffusion over the inspected specimen are needed. Most of the time, these methods involve the time dimension, i.e. the temperature evolution of the sample over the surface. For example the FFT images of the PPT characterize the temperature signal over the time of a given point at the surface of the specimen. Comparing the FFT coefficients between each pixel or from one area to the other over the surface of the sample quickly reveals defective areas that are behaving differently from sound areas. Another similar example is the TSR when used with the 1st and 2nd derivative or with polynomial coefficients. The Principal component thermography (PCT) also exploits the decomposition of the time-varying temperature signal, so does to some extent the DAC contrast. In other words, an image obtained with one of these image processing methods depends on more than one IR image of the IR sequence of images.

This presentation explains how to use the correlation coefficients as an alternative way to characterize the relative difference of the temperature evolution over the time of a sequence of IR images. The correlation coefficient technique is a well-known technique to characterize the linearity between 2 signals. Vavilov defined the correlation technique as calculating the coefficient of correlation between the temperature evolution function of each pixel and the temperature evolution function of a chosen reference point. Although the advantages and disadvantages of the correlation images have not been well explored in the context of PT and modulated thermography.

Les séminaires du LVSN ont lieu le vendredi à 11h30 dans la salle PLT-2700.


©2002-. Laboratoire de Vision et Systèmes Numériques. Tous droits réservés