Infrared Thermography for the Detection and Characterization of Buried Objects: Thesis Submitted in Fulfilment of the Requirements for the Award of ... Doctor in Engineering by Nguyen Trung Thanh

The problem of the detection and characterization of buried objects using infrared measurements of the soil surface is considered in this work that is applied to buried landmine detection. It is divided into two steps: thermal modeling of shallowly buried objects and inverse problem setting for buried object detection. The first step aims at predicting the soil temperature distribution with the presence of buried objects under natural heating conditions. For this purpose, a three-dimensional thermal model is introduced and several aspects, such as mathematical analysis, numerical methods and its applicability in practical situations, are thoroughly discussed. In the second step, the proposed thermal model and the infrared measurements are used for detecting buried objects and characterizing them by estimating their thermal and geometrical properties. Here, a three-dimensional and a simplified one-dimensional coefficient identification problem are considered. These inverse problems are severely ill-posed due to lack of spatial information in the measured data. This dissertation is mainly focused on parameterization methods, for reducing the ill-posedness of the inverse problems, and numerical algorithms for solving them. Numerical results for both simulated and real experimental data sets are demonstrated for illustrating the performance of the proposed algorithms.