基于稀疏表示的宽窄带体制气球载雷达成像理论与方法研究

Balloon-borne radar is important early warning and surveillance equipment for detecting low altitude/super-low altitude flight targets and sea targets. Enhancing the target recognition performance is a difficult problem for new balloon-borne radar. The integrate target recognition technique based on radar imaging is an effective approach to achieve the distinguishing and recognition of low altitude/super-low altitude flight targets and sea targets. However, most of the existence target recognition theory and methods which are put forward aiming at those isolated systems and models are hard to solve the problem of target recognition. In this project, by making full use of the characteristics of wideband/narrowband and low/high pulse repetition frequency coexistence system and combining with sparse representation theory, imaging theory and method of track while scan balloon-borne radar are investigated. Via breaking through some of key problems in the research contents, such as low probability of intercept waveforms design, unified sparse representation theory for target model of balloon-borne radar, wideband imaging for moving objects based on moving platform, narrowband imaging for moving objects based on moving platform, and narrowband micro-motion imaging for moving objects based on moving platform. The research can provide theoretical and technical support for new balloon-borne radar. The research of this project will have important significance and wide applications in enriching imaging theory of multi-dimension targets with complex-motion, promoting the development of target detection and recognition technology, and improving recognition performance of our country’s low altitude/super-low altitude early warning and surveillance radar.

气球载雷达是探测低空/超低飞行目标和海上目标的重要预警监视装备，提升其目标识别性能是新型气球载雷达面临的难题。基于雷达成像的综合目标识别技术是实现低空/超低飞行目标识别的有效途径。而现有目标识别理论与算法大都针对单一体制单一模式提出的，难以有效解决目标识别问题。本项目充分利用气球载雷达宽/窄带联合和低/高脉冲重复频率共存的体制特点，结合稀疏表示理论，对边扫描边跟踪体制的气球载雷达成像理论和方法展开研究。重点突破低截获概率波形设计、气球载雷达目标模型统一稀疏表示理论、运动平台运动目标宽带成像、运动平台运动目标窄带成像、运动平台窄带微动目标成像等关键问题。本项目的研究为新型气球载雷达提供理论支撑和技术支持，同时，对于丰富复杂运动目标多维度成像理论，对于促进目标识别技术的发展，对于提升我国低空/超低空预警监视雷达的识别性能等均具有重要意义和广阔应用前景。

本项目基于稀疏表示理论，研究了宽窄带体制气球载雷达理论建模、稀疏波形设计、运动目标宽带成像、窄带成像以及窄带微动目标特征提取等问题。通过四年的研究，项目按期完成了规定内容。并进一步将单基地扩展至双基地体制，研究了相应的运动目标宽带成像方法。取得的主要成果有：.（1）在球载雷达理论建模方面：完成了气球载雷达系统运动特性分析，构建了不同运动状态下的单基地雷达回波统一表示模型、双基地雷达回波模型及其特性分析、旋翼目标微动回波建模及其闪烁机理分析。.（2）在稀疏波形设计方面：针对一维与二维稀疏波形设计，分别给出了基于自相关函数以及模糊函数最大旁瓣与均值旁瓣联合约束的稀疏波形设计方法，所提方法具有计算量小、易实现等优点。.（3）在运动目标宽带成像方面：首先研究了波形稀疏条件下目标高精度运行补偿方法；基于此，从提升算法重构效率与鲁棒性角度出发，提出多种信号稀疏条件下的一维距离像重构以及二维ISAR成像算法，通过仿真数据以及实测数据验证了所提算法具有稳健、快速的高分辨成像性能。.（4）在运动目标窄带成像方面：针对欠采样和遮挡等不利因素，研究了基于窄带的自旋目标高分辨成像方法，并分析了其高分辨成像机理。此外，进一步提出基于分布式匹配稀疏表示模型的宽带旋转部件快速超分辨成像方法，实现了欠采样以及低信噪比下宽带自旋目标高分辨成像。.（5）在窄带微动目标特征提取方面：首先，提出一种变步长目标微动参数提取方法，能够通过改变搜索步长实现对叶片转速和尺寸的快速估计。其次，进一步提出基于谱聚集度指标的微动特征提取方法，通过奇异值比谱法估计转动频率，配合逐次消去法以及叶片初相之间的关系实现了低信噪比下旋翼长度、初相和叶片数等参数估计。.（6）在参数化稀疏表征成像方面:通过构建机动目标参数化稀疏表征模型，将成像稀疏解与稀疏基参数求解转化为联合优化问题，利用遗传算法实现对参数的快速搜索，实现了稀疏孔径下机动目标的快速成像。