Improvements in GPR-SAR imaging focusing and detection capabilities of UAV-mounted GPR systems
Improvements in GPR-SAR imaging focusing and detection capabilities of UAV-mounted GPR systems
复制标题
DOI：
10.1016/j.isprsjprs.2022.04.014
复制
发表时间：
2022-07
影响因子：
12.7
通讯作者：
M. García-Fernández;G. Álvarez-Narciandi;Yuri Álvarez López;F. Las-Heras Andrés
中科院分区：
文献类型：
--
作者：
M. García-Fernández;G. Álvarez-Narciandi;Yuri Álvarez López;F. Las-Heras Andrés
近年来，人们对开发机载探地雷达（GPR）系统的研究兴趣日益浓厚，该系统可安全快速地检测地雷和简易爆炸装置（IED）等埋藏威胁。与金属探测器或磁力计等其他传感器相比，探地雷达能够检测金属和非金属目标。此外，使用超宽带（UWB）射频硬件可以检索地下土壤的高分辨率图像，从而提高检测能力。机载探地雷达系统受到不确定性的影响，需要对其进行表征和纠正。本贡献重点关注以下三个问题：（i）探地雷达合成孔径雷达（GPR-SAR）图像中高度信息的影响，确定哪些传感器提供更好质量的图像； (ii) 由于无人机上 UWB 天线的小倾斜而导致沿航迹轴成像目标的畸变，提出了一种联合配准技术来纠正该畸变； (iii) SVD 滤波技术的修订，评估对不同特征图像进行分类的标准。对这些问题的分析以及对所提出的校正方法的验证是使用在已实施的机载探地雷达系统验证活动期间进行的测量来进行的。为此，选择了两种具有不同土壤特性的验证方案。其中埋有不同形状和大小的金属和非金属靶材。在第一种情况（潮湿的壤土）中，检测能力从 50%（检测到 10 个目标中的 5 个）提高到 80%。对于第二个验证场景（壤土，但比第一个场景湿度低），不仅实现了 100% 的检测，而且通过本贡献中提出的改进获得的 GPR-SAR 图像中的目标也清晰成像。
In recent years there has been an increasing research interest in the development of airborne-based Ground Penetrating Radar (GPR) systems for safe and fast detection of buried threats such as landmines and Improvised Explosive Devices (IEDs). Compared to other sensors such as metal detectors or magnetometers, GPR is able to detect either metallic and non-metallic targets. Besides, the use of Ultra Wide Band (UWB) radiofrequency hardware allows retrieving high resolution images of the subsoil, thus improving the detection capabilities. Airborne-based GPR systems are affected by uncertainties that need to be characterized and corrected. This contribution focuses on the following three issues: (i) influence of the height information in the GPR-Synthetic Aperture Radar (GPR-SAR) images, determining which sensors provide better quality images; (ii) distortion of the imaged targets in along-track axis due to small tilting of the UWB antennas on board the UAV, proposing a co-registration technique to correct it; and (iii) revision of SVD filtering techniques, assessing a criterion to classify the different eigenimages. The analysis of these issues, as well as the validation of the proposed correction methods, are conducted using measurements taken during validation campaigns of the implemented airborne-based GPR system. For this purpose, two validation scenarios with different soil characteristics have been selected. Metallic and non-metallic targets with different shapes and sizes have been buried in them. In the case of the first scenario (wet loamy soil), detection capabilities improved from 50% (5 out of 10 targets were detected) to 80%. For the second validation scenario (loamy soil, but less humid than the first scenario), not only a 100% detection is achieved, but also the targets are clearly imaged in the GPR-SAR images obtained with the improvements presented in this contribution.