This paper deals with active time-harmonic infrared thermography applied to the detection of defects inside thin plates. We propose a method to post-process raw thermograms based on the computation of topological derivatives which will produce much sharper images (namely, where contrast is highly enhanced) than the original thermograms. The reconstruction algorithm does not need information about the number of defects, nor the size or position. A collection of numerical experiments illustrates that the algorithm is highly robust against measurement errors in the thermograms, giving a good approximation of the shape, position and number of defects without the need of an iterative process.

中文翻译：

---

拓扑导数在红外时谐波热谱图后期处理中的应用

本文将主动时谐波红外热成像技术应用于薄板内部缺陷的检测。我们提出了一种基于拓扑导数计算的原始热谱图后处理方法，该方法将比原始热谱图产生更清晰的图像（即对比度大大增强）。重建算法不需要有关缺陷数量，大小或位置的信息。一组数值实验表明，该算法对热分析图中的测量误差具有很高的鲁棒性，无需迭代过程即可很好地近似缺陷的形状，位置和数量。