Automatic Target Recognition for Hyperfine Resolution Synthetic Aperture Radar

本研究的目的是了解在最近开发的精细分辨率合成孔径雷达（SAR）系统的范围内，新开发的自动目标识别（ATR）概念的可能极限。SAR作为一种宝贵的遥感能力，自其早期成立以来已经取得了显著的发展，由于其全天候、昼夜和远程能力，与其他遥感方式相比，它是独一无二的。近年来，SAR系统已经能够传输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，具有在场景中检测和分类感兴趣目标的巨大潜力。近年来，由于光线追踪和时域有限差分（FDTD）软件的新功能以及越来越多的开源仿真工具可供研究人员使用，SAR仿真领域也取得了进展。其中包括由Stefan Auger[1]和GPRMax开发的RaySAR工具[2-3]。拟议的研究将寻求利用开源软件工具开发一个能够提供高保真SAR图像的模拟环境，以馈送到ATR分类技术中

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