无人机SAR数据低比特量化及其复杂度分析

1)合肥学院计算机科学与技术系,安徽合肥 230601; 2)太原卫星发射中心技术部,山西太原 030027; 3)西安电子科技大学微电子学院,陕西西安 710071

信息处理技术; 雷达工程; 无人机; 合成孔径雷达; 低比特量化; 单频阈值

Low-bit quantization for UAV SAR data and its complexity analysis
LI Zhengmao1, CHEN Daqing2, and LIU Maliang3

1)Department of Computer Science and Technology, Hefei University, Hefei 230601, Anhui Province, P.R.China2)Technical Department, Taiyuan Satellite Launch Center, Taiyuan 030027, Shanxi Province, P.R.China3)School of Microelectronics, Xidian University, Xi'an 710071, Shaanxi Province, P.R.China

information processing technique; radar engineering; unmanned aerial vehicle(UAV); synthetic aperture radar(SAR); low-bit quantization; single-frequency threshold

DOI: 10.3724/SP.J.1249.2019.05503

备注

无人机(unmanned aerial vehicle, UAV)平台由于尺寸与载荷的限制,仅能提供有限的硬件计算资源.如何利用尽可能低的计算复杂度来获得高质量的合成孔径雷达(synthetic aperture radar, SAR)成像,是基于UAV SAR系统设计面临的一个重要问题.分析现有算法的缺陷,利用不同的量化方法对SAR原始回波数据进行低比特量化,针对低比特量化SAR数据提出基于单频阈值的成像质量改善方法.在此基础上,研究不同量化策略下SAR成像匹配滤波处理所需的计算复杂度,并定量评估了相应的成像质量.研究结果可为基于UAV平台的SAR成像应用中计算复杂度与成像质量之间的取舍提供参考.

Due to the limitation of size and load, an unmanned aerial vehicle(UAV)can only provide limited hardware computing resources. How to obtain high-quality synthetic aperture radar(SAR)imaging with the lowest computational complexity is an important problem in UAV-based SAR system design. In this paper, the disadvantages of existing algorithms are analyzed. Different quantization approaches are used to quantify the raw data of SAR in low bits. An image quality improvement method based on a single frequency threshold is proposed for low-bit quantized SAR data. On this basis, the computational complexity of SAR imaging matched filtering processing under different quantization strategies is studied, and the corresponding imaging quality is quantitively evaluated. This paper provides a reference for achieving the tradeoff between computational complexity and imaging quality in SAR imaging application based on UAV platform.

·