One of our industrial partners, Assek Technologie, is interested in developing a technique that would improve the
drying process of wood floor in basements after flooding. In order to optimize the procedure, the floor structure and the
damaged (wet) area extent must first be determined with minimum intrusion (minimum or no dismantling). The present
study presents the use of infrared thermography to reveal the structure of (flooded) wood floors. The procedure involves
opening holes in the floor. Injecting some hot air through those holes reveals the framing structure even if the floor is
covered by vinyl or ceramic tiles. This study indicates that thermal imaging can also be used as a tool to validate the
decontamination process after drying. Thermal images were obtained on small-scale models and in a demonstration
IR-View, is a free and open source Matlab software that was released in 1998 at the Computer Vision and Systems Laboratory (CVSL) at Université Laval, Canada, as an answer to many common and recurrent needs in Infrared thermography. IR-View has proven to be a useful tool at CVSL for the past 10 years. The software by itself and/or its concept and functions may be of interest for other laboratories and companies working in research in the IR NDT field. This article describes the functions and processing techniques integrated to IR-View, freely downloadable under the GNU license at http://mivim.gel.ulaval.ca. Demonstration of IR-View functionalities will also be done during the DSS08 SPIE Defense and Security Symposium.
In non-destructive testing by Infrared Thermography it is usually needed to locate defects and region of interests suspected to contain defects. The defects cannot always be observed directly from one single IR image taken at a single given time t. Thus, in the case of pulsed thermography, direct course techniques as the Fourier transform process the information of many images recorded for a given duration into one resulting image. Another way to compile the temporal information of a sequence of images into a single one is to compute a correlation image. This paper details an approach to use a statistical correlation operator to help improving defect detection in pulsed infrared thermography.
This paper presents a summary of recent research activities carried out at our laboratory in the field of Infrared Thermography for Nondestructive Evaluation (TNDE). First, we explore the latest developments in signal improvement. We describe three approaches: multiple pulse stimulation; the use of Synthetic Data for de-noising of the signal; and a new approach derived from the Fourier diffusion equation called the Differentiated Absolute Contrast method (DAC). Secondly, we examine the advances carried out in inverse solutions. We describe the use of the Wavelet Transform to manage pulsed thermographic data, and we present a summary on Neural Networks for TNDE. Finally, we look at the problem of complex geometry inspection. In this case, due to surface shape, heat variations might be incorrectly identified as flaws. We describe the Shape-from-Heating approach and we propose some potential research avenues to deal with this problem.