基于GPU的SAR成像算法设计与实现
发布时间：2019-04-09 17:17
【摘要】：合成孔径雷达(Synthetic Aperture Radar,简称SAR)是近几十年发展起来的新型雷达,它可以不受光照、天气的影响实现测绘区域内的二维高分辨率成像,SAR被广泛应用在军事、工业,农业、海洋探测、地质探测等领域。要对测绘带进行成像,需要处理大量的回波数据,SAR成像的实质就是利用信号处理的方法来处理这些数据,从而获取满足要求的高分辨率图像。SAR成像算法复杂,涉及大量的运算,传统通过CPU进行SAR成像算法实现在实时性上越来越乏力,于是人们不断地探求信号处理新技术来满足实时性要求。自NVIDIA推出GPU(Graphic Processing Unit)以来,经过十几年的快速发展,GPU的使用范围已不再局限于进行图像处理与显示,尤其是2007年NVIDIA推出基于CUDA(Compute Unified Device Archetecture)的新架构GPU,人们初步体验到了GPU做高速并行计算的强大能力,在全球范围内掀起了研究GPU算法提速科学运算的潮流。SAR的工作者们也察觉到了这一技术的优势,尝试将SAR成像算法移植到GPU上,从而满足自身需要,缩短成像时间。本文主要研究了GPU在SAR成像方面的应用,设计了基于GPU的SAR成像算法,利用GPU做并行计算的优势,缩短SAR成像时间,提高仿真系统的性能。本文主要涉及到的内容有:介绍了SAR成像原理,详细的说明了如何通过脉冲压缩技术来获取高分辨率雷达图像,给出了正侧视模型下的二维频域成像算法,给出了适用于前斜视与俯冲前斜的成像算法,并对这三种模型做了比较。介绍了GPU的技术特点,设计了基于GPU的SAR成像算法。设计了一个基于GPU的SAR成像仿真系统,该系统在主控界面输入雷达系统参数并启动成像算法,最后将成像结果显示在界面上。本论文的主要贡献和创新是:(1)设计了方位向空变校正的前斜视SAR成像算法。(2)设计了方位向空变校正的俯冲段SAR成像算法。(3)设计了基于GPU的正侧视、前斜视、俯冲段SAR成像算法。(4)设计了基于GPU的SAR成像结果测试系统。
[Abstract]:Synthetic Aperture Radar (Synthetic Aperture Radar,) is a new type of radar developed in recent decades. It can realize two-dimensional high-resolution imaging in mapping area without the influence of light and weather. SAR is widely used in military, industry and agriculture. Ocean exploration, geological exploration and other fields. It is necessary to process a large amount of echo data in order to image the mapping tape. The essence of SAR imaging is to process these data by means of signal processing, so as to obtain high-resolution images that meet the requirements. Involving a large number of operations, the traditional SAR imaging algorithm through CPU is less and less powerful in real-time performance, so people continue to explore new signal processing techniques to meet the real-time requirements. Since the introduction of GPU (Graphic Processing Unit) by NVIDIA, after more than ten years of rapid development, the scope of use of GPU is no longer limited to image processing and display, especially in 2007, NVIDIA launched a new CUDA (Compute Unified Device Archetecture)-based architecture, GPU, People have initially experienced the powerful ability of GPU to do high-speed parallel computing, and set off the trend of research on speed-up and scientific computing of GPU algorithm on a global scale. Sar workers are also aware of the advantages of this technology. This paper attempts to transplant the SAR imaging algorithm to the GPU, so as to meet its own needs and shorten the imaging time. This paper mainly studies the application of GPU in SAR imaging, designs the SAR imaging algorithm based on GPU, makes use of the advantage of GPU to do parallel computation, shortens the SAR imaging time, and improves the performance of the simulation system. The main contents of this paper are as follows: the principle of SAR imaging is introduced, how to obtain high-resolution radar image by pulse compression technique is explained in detail, and the two-dimensional frequency domain imaging algorithm based on forward and side-looking model is given. The imaging algorithms for forward strabismus and subduction forward squint are presented, and the three models are compared. The technical characteristics of GPU are introduced, and the SAR imaging algorithm based on GPU is designed. A SAR imaging simulation system based on GPU was designed. The system inputted the parameters of radar system at the main control interface and started the imaging algorithm. Finally, the imaging results were displayed on the interface. The main contributions and innovations of this thesis are as follows: (1) the forward squint SAR imaging algorithm for azimuth space variation correction is designed. (2) the SAR imaging algorithm for subduction segment is designed based on azimuth space variation correction. (3) the forward side view and forward squint view based on GPU are designed. Subduction segment SAR imaging algorithm. (4) A testing system of SAR imaging results based on GPU is designed.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
[Abstract]:Synthetic Aperture Radar (Synthetic Aperture Radar,) is a new type of radar developed in recent decades. It can realize two-dimensional high-resolution imaging in mapping area without the influence of light and weather. SAR is widely used in military, industry and agriculture. Ocean exploration, geological exploration and other fields. It is necessary to process a large amount of echo data in order to image the mapping tape. The essence of SAR imaging is to process these data by means of signal processing, so as to obtain high-resolution images that meet the requirements. Involving a large number of operations, the traditional SAR imaging algorithm through CPU is less and less powerful in real-time performance, so people continue to explore new signal processing techniques to meet the real-time requirements. Since the introduction of GPU (Graphic Processing Unit) by NVIDIA, after more than ten years of rapid development, the scope of use of GPU is no longer limited to image processing and display, especially in 2007, NVIDIA launched a new CUDA (Compute Unified Device Archetecture)-based architecture, GPU, People have initially experienced the powerful ability of GPU to do high-speed parallel computing, and set off the trend of research on speed-up and scientific computing of GPU algorithm on a global scale. Sar workers are also aware of the advantages of this technology. This paper attempts to transplant the SAR imaging algorithm to the GPU, so as to meet its own needs and shorten the imaging time. This paper mainly studies the application of GPU in SAR imaging, designs the SAR imaging algorithm based on GPU, makes use of the advantage of GPU to do parallel computation, shortens the SAR imaging time, and improves the performance of the simulation system. The main contents of this paper are as follows: the principle of SAR imaging is introduced, how to obtain high-resolution radar image by pulse compression technique is explained in detail, and the two-dimensional frequency domain imaging algorithm based on forward and side-looking model is given. The imaging algorithms for forward strabismus and subduction forward squint are presented, and the three models are compared. The technical characteristics of GPU are introduced, and the SAR imaging algorithm based on GPU is designed. A SAR imaging simulation system based on GPU was designed. The system inputted the parameters of radar system at the main control interface and started the imaging algorithm. Finally, the imaging results were displayed on the interface. The main contributions and innovations of this thesis are as follows: (1) the forward squint SAR imaging algorithm for azimuth space variation correction is designed. (2) the SAR imaging algorithm for subduction segment is designed based on azimuth space variation correction. (3) the forward side view and forward squint view based on GPU are designed. Subduction segment SAR imaging algorithm. (4) A testing system of SAR imaging results based on GPU is designed.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
【相似文献】
相关期刊论文 前10条
1 李述为;晏磊;杨绍文;高梅国;傅雄军;孙华波;;一种雷达穿墙成像算法[J];系统工程与电子技术;2009年03期
2 张二磊;张建奇;;合成孔径激光雷达成像算法[J];电子科技;2009年11期
3 严晓浪,陈晔,史峥,马s，
本文编号：2455381
本文链接：https://www.wllwen.com/kejilunwen/wltx/2455381.html
