고기동 비협조 이동표적에 대한 ISAR 영상 형성 연구

Abstract

Usually, the inverse synthetic aperture radar (ISAR) imaging problem is more challenging than synthetic aperture radar (SAR) imaging. In SAR, the well-controlled radar platform is usually known, whereas the target motion is usually unknown in ISAR system. Therefore, how to deal with unknown target motion is a critical issue in ISAR imaging. In this dissertation, we address major problems, caused by the non-cooperative characteristics of the target for ISAR imaging. When a target undergoes complex three-dimensional (3D) motions, focused ISAR images cannot be obtained using any motion compensation (MOCOM) algorithms. To address this problem, in chapter 2, we propose a method to efficiently determine the suitable frame time and length for ISAR imaging, which exploits phase nonlinearity and discrete polynomial phase transforms. Once suitable frame times are effectively determined, we should perform MOCOM to remove undesired target motions. However, conventional ISAR MOCOM schemes are not optimal, in terms of reconstructed image and/or computation time. Motivated by the problems of the conventional schemes, in Chapter 3, we introduce an efficient autofocus chain for ISAR imaging of non-cooperative moving target. From the experimental results using real data sets, we can conclude that the proposed method is highly efficient to remove undesired motion errors, in terms of both image quality and computational efficiency. Focused ISAR image can be obtained in range-Doppler (RD) domain after applying frame-selection scheme and MOCOM. However, RD image is inefficient for target classification, due to the variable Doppler scaling factor related to target’s own rotational motion. Therefore, for more effective use of ISAR image, ISAR images should be rescaled in the homogeneous range and cross-range domain by estimating rotation velocity of a non-cooperative targets. In Chapter 4, we propose particle swarm optimization combined with exhaustive search method (PSO-ESM) for ISAR cross-range scaling (CRS). Robust scatterers against angular scintillation are extracted using scale invariant feature transform, and locations of the extracted scatterers are applied to PSO-ESM. We should note that the PSO-ESM enables to perform robust CRS by joint estimation of rotation center and velocity. Meanwhile, all aforementioned studies are related to monostatic radar configurations where a transmitter and a receiver are collocated. However, monostatic ISAR imaging suffers from some limitations, including geometrical issues and imaging problems with regard to stealthy targets. In a monostatic radar configuration, the aspect angle of a target does not change relative to the radar when the target moves along the radar line of sight (LOS). In this case, scatterers distributed on the target cannot be separated in the cross-range direction. In addition, the imaging of a stealthy target is difficult in this configuration because a stealthy target is designed to reflect electromagnetic energy to directions other than that of the radar, yielding a decrease in the signal-to-noise ratio (SNR) of a received radar signal. To overcome these problems, a bistatic radar configuration where the transmitter and receiver are spatially separated has been considered for ISAR imaging. Because the bistatic ISAR (Bi-ISAR) geometry provides adequate look-angle diversity of the target, rotation of the target with respect to the radar is ensured for the acquisition of the desired cross-range resolution. In addition, imaging of a stealthy target can be easily achieved using bistatic radars because the SNR of the received signal in the bistatc configurations increases compared with a monostatic radar. Therefore, recently, studies related to Bi-ISAR imaging has been widely studied. In Chapter 5, we introduce a bistatic inverse synthetic aperture radar (Bi-ISAR) CRS method to more effectively use Bi-ISAR images in their applications. We note that monostatic ISAR CRS method, presented in Chapter 4, cannot be applied in Bi-ISAR configurations. For this, we propose a method to estimate the effective rotation velocity of a target in a Bi-ISAR imaging system, and restore a linear-geometry distortion that yields a sheared shape of Bi-ISAR images. In the simulations, we observed that our proposed method is capable of performing robust and precise Bi-ISAR CRS. In addition, monostatic MOCOM approaches fails to provide focused ISAR images in bistatic radar configuration. To address this problem, in Chapter 6, a new MOCOM framework for Bi-ISAR imaging is proposed. In the proposed framework, the undesired motion errors can be correctly removed in Bi-ISAR systems, yielding focused Bi-ISAR images. In addition, restoration of the sheared Bi-ISAR images is achieved in the proposed framework. In the simulations and experiments, we found that the proposed Bi-ISAR MOCOM framework can provide high-quality Bi-ISAR images.