One dimensional cross-range imaging and methods to improve the resolution of low resolution radar targets

One dimensional (1D) images can be used to characterize the major features of the target and serve for the purpose of automatic target recognition (ATR). For conventional low resolution radar, 1D range profiles cannot be measured, but 1D cross-range images can be obtained by processing a sequence of radar echoes. There have been few papers in the literature discussing how to obtain 1D cross-range images, but the image resolutions obtained are low, and even some wrong images were obtained because of the following two reasons: one is that translational motion compensation (TMC) is not correct; the other is that 3D and nonuniform rotational motions are not considered. In this paper we propose new methods to tackle these two problems. First we use the minimum entropy method to compensate the translational motion. Second, based on the observation that the scatterer sub-echoes are not sinusoidal and should be approximated as linear frequency modulation (LFM) signals after TMC due to the 3D and nonuniform rotation of the target, we propose an extended RELAX high resolution method to estimate the multicomponent LFM signals, and obtain dynamic instantaneous 1D cross-range images. The methods are evaluated using both simulated data and real radar data