This paper focuses on the synthetic aperture radar (SAR)imaging of space-sparse targets such as ships on the sea,and proposes a method of targets separation and imaging of sparse scene based on cluster result of range profile peaks.Firstly,wavelet de-noising algorithm is used to preprocess the original echo,and then the range profile at different viewing positions can be obtained by range compression and range migration correction.Peaks of the range profi les can be detected by the fast peak detection algorithm based on second order difference operator.Targets with sparse energy intervals can be imaged through azimuth compression after clustering of peaks in range dimension.What’s more,targets without coupling in range energy interval and direction synthetic aperture time can be imaged through azimuth compression after clustering of peaks both in range and direction dimension.Lastly,the effectiveness of the proposed method is validated by simulations.Results of experiment demonstrate that space-sparse targets such as ships can be imaged separately and completely with a small computation in azimuth compression, and the images are more beneficial for target recognition.%针对海面舰船等具有一定空间稀疏性的合成孔径雷达成像场景，提出了一种稀疏场景目标的距离像峰值聚类分割成像方法。首先采用小波降噪算法对 SAR 原始回波数据进行预处理，通过距离压缩和距离徙动校正获得不同观测位置的距离像，利用基于二阶差分算子的快速峰值检测算法检测距离像峰值点，对峰值检测结果距离维聚类后方位向成像，实现了距离向能量区间稀疏目标的分割成像；对峰值检测结果距离-方位二维聚类后方位向成像，实现了距离向能量区间与方位向合成孔径时间无耦合稀疏目标的分割成像。仿真结果表明，对海面舰船等具有空间稀疏性的成像场景，所提方法能够实现目标的有效分割成像，不仅在完整保留目标回波信息的同时大幅度地降低了方位向压缩的运算量，而且分割成像结果更有利于目标的快速识别。
展开▼
