Automatic Target Recognition for Hyperfine Resolution Synthetic Aperture Radar

这项研究的目的是在最近发展的高分辨率合成孔径雷达(SAR)系统的范围内，了解新发展的自动目标识别(ATR)概念的可能限度。自早期建立以来，合成孔径雷达作为一种无价的遥感能力得到了长足的发展，由于其全天候、全天候和远距离能力，与其他遥感方式相比是独立的。近年来，合成孔径雷达系统已经能够发送和接收越来越宽的带宽信号，就可以形成的图像而言，这产生了更精细的图像，能够产生令人难以置信的详细图像，具有很大的潜力来检测和分类场景中的感兴趣目标。随之而来的挑战是，由于光线跟踪和时域有限差分(FDTD)软件的新功能，合成孔径雷达(SAR)仿真领域近年来也取得了进展，越来越多的开源仿真工具可供研究人员使用。其中包括Stefan Auger开发的RaySAR工具[1]和GPRMax[2-3]。拟议的研究将寻求利用开源软件工具来开发一个模拟环境，该环境能够提供高保真的SAR图像，以提供给ATR分类技术

The goal of this research is to understand, within the bound of recently developed fine resolution Synthetic Aperture Radar (SAR) systems, the limits of the possible of newly developed Automatic Target Recognition (ATR) concepts. SAR has developed significantly since its early inception as an invaluable remote sensing capability that stands alone compared to other sensing modalities due to its all weather, day/night and long range capabilities. In recent years SAR systems have been able to transmit and receive wider and wider bandwidth signals, in terms of the images that can be formed this produced a much finer image that is able to produce incredibly detailed imagery that has substantial potential to both detect and classify targets of interest within the scene. The resulting challenge from this is the The area of SAR simulation has also progressed in recent years due to new capabilities in ray tracing and Finite-Difference Time-Domain (FDTD) software and more and more open source simulation tools have become available to researchers. These include RaySAR tool developed by Stefan Auger [1] and the GPRMax [2-3]. The proposed research would look to utilize open source software tools to develop a simulation environment that is able to provide high fidelity SAR imagery to feed into ATR classification techniques