穿墙探测：移动物体的反演与成像

Through wall detection has many important applications, for example, searching buried human beings in earthquake debris and locating terrorists and hostages inside a building. Using ultra-wind band (UWB) radar to detect and image moving targets (including human beings) behind walls is a cutting-edge research area. This proposal concerns the detection and imaging for multiple visually obscured moving targets. Based on parallel numerical simulation for electromagnetic scattering for UWB radar, we shall development time-domain direct methods to reconstruct the trace of the targets. Frequency domain analysis will be used to detect human respiration and heart beating. Finally, full-wave inversion will be developed to obtain the electric permittivity and conductivity. Successful implementation of the project will enhance the usage of UWB radar in through wall detection, advance the interdisciplinary research of inverse scattering and engineering applications, and provide information for UWB radar design.

穿墙探测在现实生活中有很多重要的应用，例如地震救援时探测被废墟掩埋的人员，反恐行动中确定隐藏在建筑物中恐怖分子和人质的位置等等。利用超宽带雷达对建筑物中移动的目标（包括有生命体征如呼吸和心跳的人员）进行探测，成像和反演是国际上工程反问题领域的研究重点。本项目旨在研究移动多目标的探测和成像。在开发超宽带雷达电磁散射的并行数值模拟基础上，利用反散射领域最新的成果，发展时域直接成像方法来重构目标的运动轨迹，进而应用频域分析技术来侦测目标是否具有生命体征，最终利用全波反演得到介电常数和电导率，完成对移动目标的准确刻画。项目的成功实施将提高超宽带雷达在穿墙探测中的作用，促进反散射理论与工程实际的结合，并对超宽带雷达的设计提供重要的指导。

本项目重点研究移动物体探测与成像的数学理论和算法。其实际应用的背景为穿墙探测。发展了新的处理部分数据反散射问题的直接方法，以及贝叶斯反问题方法。.实现了穿墙探测的数值模拟与探测算法。为用较少数据进行目标的探测成像提供了重要的研究路径和先期结果。