Aliasing-free high resolution imaging of fast rotating targets with narrowband radar
来源期刊:中南大学学报(英文版)2014年第5期
论文作者:YOU Peng(游鹏) LIU Zhen(刘振) WEI Xi-zhang(魏玺章) WANG Hong-qiang(王宏强) LI Xiang(黎湘)
文章页码:1842 - 1851
Key words:narrowband radar imaging; fast rotating; compressed sensing; random pulse repetitive interval; aliasing reduction
Abstract: Narrowband radar has been successfully used for high resolution imaging of fast rotating targets by exploiting their micro-motion features. In some practical situations, however, the target image may suffer from aliasing due to the fixed pulse repetition interval (PRI) of traditional radar scheme. In this work, the random PRI signal associated with compressed sensing (CS) theory was introduced for aliasing reduction to obtain high resolution images of fast rotating targets. To circumvent the large-scale dictionary and high computational complexity problem arising from direct application of CS theory, the low resolution image was firstly generated by applying a modified generalized Radon transform on the time-frequency domain, and then the dictionary was scaled down by random undersampling as well as the atoms extraction according to those strong scattering areas of the low resolution image. The scale-down-dictionary CS (SDD-CS) processing scheme was detailed and simulation results show that the SDD-CS scheme for narrowband radar can achieve preferable images with no aliasing as well as acceptable computational cost.
YOU Peng(游鹏), LIU Zhen(刘振), WEI Xi-zhang(魏玺章), WANG Hong-qiang(王宏强), LI Xiang(黎湘)
(School of Electronic Science and Engineering,
National University of Defense Technology, Changsha 410073, China)
Abstract:Narrowband radar has been successfully used for high resolution imaging of fast rotating targets by exploiting their micro-motion features. In some practical situations, however, the target image may suffer from aliasing due to the fixed pulse repetition interval (PRI) of traditional radar scheme. In this work, the random PRI signal associated with compressed sensing (CS) theory was introduced for aliasing reduction to obtain high resolution images of fast rotating targets. To circumvent the large-scale dictionary and high computational complexity problem arising from direct application of CS theory, the low resolution image was firstly generated by applying a modified generalized Radon transform on the time-frequency domain, and then the dictionary was scaled down by random undersampling as well as the atoms extraction according to those strong scattering areas of the low resolution image. The scale-down-dictionary CS (SDD-CS) processing scheme was detailed and simulation results show that the SDD-CS scheme for narrowband radar can achieve preferable images with no aliasing as well as acceptable computational cost.
Key words:narrowband radar imaging; fast rotating; compressed sensing; random pulse repetitive interval; aliasing reduction