摘要
现代谱估计技术被用于雷达成像中来提高图像分辨率和目标特征提取,但其缺点是需耗时的二维谱峰搜索,且由于采用码赛克技术,算法性能受小样本数限制.文中将成像高分辨算法中的两维参数估计问题转化为两个独立的一维问题,然后利用ESPRIT方法或MUSIC方法求得每维参数估计,这种方法无需两维谱峰搜索,算法性能在小样本时有很大改善,计算机仿真结果表明了文中算法的有效性.
Modern spectral estimation methods have recently been used to obtain a higher resolution in radar imaging and features extraction of a target. But they are time consuming while searching in the 2 D space, and their performance is limited by the available snapshots in mosaicing strategy. In this paper, the 2 D parameters of the scatterers are obtained via two separate 1 D estimation procedures, where ESPRIT or MUSIC methods are used to get the 1 D estimates. As a result, the need of searching in 2 D space is avoided with an improved resolution at low SNR and small short snapshots. Simulation results show the effectiveness of this algorithm.
出处
《西安电子科技大学学报》
EI
CAS
CSCD
北大核心
1999年第6期737-742,共6页
Journal of Xidian University
基金
国家自然科学基金
关键词
雷达成像
高分辨技术
算法
分辨率
radar imaging
high resolution
parameters estimation
moving average