摘要

Infrared thermography allows contactless non-destructive testing of objects based on their thermal behavior. Quantitative inspection of an object aims to characterize its internal defects by estimating their size and their depth. In the field, a one-dimensional thermal model has been used for depth estimation. Unfortunately, the methods based on this model become inaccurate when the inspected object has a high thermal diffusivity, a complex shape, or when the defects, like corrosion, have a complex geometry. For such cases, a 3D formulation of the problem is needed. In this paper, we consider the defect characterization as an inverse geometry problem and we propose a new method: the rear surface reconstruction by temporal tracking of the thermal front. The idea is to follow the thermal front while it propagates inside the object. Referring to the duality time-depth, at every moment, the penetration depth of the thermal front can be estimated. As soon as the thermal front reaches the rear surface, a temperature change will be noticeable on the frontal surface. It is then possible to update the internal geometry of the object at each time step in such a way that the difference between the theoretical temperature, obtained by a 3D solver, and the experimental temperature, recorded by an infrared camera, is minimized. The proposed method shows accurate results and can address situations involving rear surfaces with complex geometry and objects with high thermal diffusivity and a complex shape.

  • 出版日期2016-12