Study Of New Methods For ISAR Imaging

Inverse Synthetic Aperture Radar (ISAR) is an important kind of high resolution radar. It has the characteristic of day/night, all-weather, far range, and can enhance radar's information inquisition capability greatly, and thus has great value in both civilian and military application. With the increasing demand, ISAR imaging has attract more and more attentions. In order to fully exert signal processing influence and improve spatial target imaging capability of the radar, for some aspects of ISAR imaging, measures such as conceptual analysis, theoretical derivation and data demonstration are studied in this dissertation. The main work of this dissertation is as follows:1. In the inverse synthetic aperture radar (ISAR) imaging, the image will be greatly degraded by the nonuniform rotation of the target. To address this issue, an efficient method for ISAR imaging is proposed. In the proposed method, a time frequency analysis is firstly applied to single range unit cross-range echo data. Secondly, according to the relationship between the rotating speed and the scaling factor, the referenced signal is constructed by scaling the frequency distribution of the middle time with many different scaling factors. Thirdly, the curve of the rotating speed is estimated by figuring out the correlation coefficient between the referenced signal and the frequency distribution of other moments. Then, an interpolation is executed in azimuth to compensate for the effects of changes in rotation speed. Finally, the well focused ISAR image of the target is obtained. As all the changes of the target rotating speed are exactly estimated and compensated, the proposed approach is applicable to all the targets however the rotating speed changes. Moreover, it is computational efficient and can be used in real time processing. The results of simulation and real data are provided to confirm the validity of the proposed approach.2. With the improvement of resolution in the inverse synthetic aperture radar (ISAR) imaging, the effect of the rotational motion of the target is much larger than ever before. A novel subimage fusion method is proposed to compensate for the large rotational motion in high-resolution ISAR imaging. By dividing the large rotational motion into several small ones, the subimage formation can be realized easily with current ISAR imaging algorithms. At the same time, with the increasing of the number of the subimages, the change of the aspect angle during the coherent processing interval can be very large, and high resolution can be achieved. Moreover, because the whole processing is achieved with Fast Fourier Transforms and vector-multiplication operations, the proposed approach is computationally efficient and has no interpolation operation. Numerical examples are provided to demonstrate the performance of the proposed approach.3. Ground-based radar is a powerful tool for monitoring space debris. Imaging of the space debris with dimensions smaller than the range resolution is of great significance. Based on the observation that space debris usually follows a simple spin motion around its major axis, and by analyzing the signal model of the target, two novel methods for radar imaging of space debris via sparse signal representation are proposed, that is novel single-range image fusion (SRIF) method and the imaging method via sparse signal representation. On the assumption that the target consists of isolated isotropic scattering centers, a 2-D image can be obtained using single-range unit cross-range echo data, which have a theoretical resolution of a quarter of a wavelength. The first proposed approach is computationally efficient especially for smaller-sized targets, as the total rotational angle is divided into four small sections, allowing for easier processing of each region. Moreover, it can be used to directly obtain an image in Cartesian coordinates, unlike current algorithms where images are obtained in a polar format, requiring reformatting to the Cartesian grid. The validity of the first proposed approach is confirmed using numerical simulations. In the second proposed method, the imaging problem is converted to the problem of sparse signal decomposition. Firstly, an overcomplete dictionary is constructed by elements which represent the scattering properties of each point. Then, a sparse representation of the single range unit radar echo is resolved by the current method. Finally, the 2D high resolution image of the target is obtained by rearranging the resolved result. Furthermore, the validity of the second proposed approach is also confirmed using numerical simulations.4. Frequency-stepped chirp signal can simplify the designation of radar system. However, it has a shortcoming of Doppler ambiguity for high-speed moving targets. Therefore, it is of great significance to study how to increase its equivalent pulse repeat frequency. The back scattering field of the ISAR target has strong sparsity, that is to say, most energy is contributed merely by a few scattering centers. Hence, based on the theory of the sparse signal representation, a novel method for ISAR imaging via sparse frequency-stepped chirp signals is proposed by analyzing the signal model of the target. In the proposed method, partial of sub-pulses of the frequency-stepped chirp signal is randomly selected to transmit, and then the 2D high-resolution image of the target can be constructed by sparse signal decomposition. At the cost of computational resources, the method can effectively resolve the problem of Doppler ambiguity, decrease the sidelobes and obtain a super-resolution image. Furthermore, the validity of the proposed approach is confirmed by the results of numerical simulations and real data.5. Based on compressive sampling theory, a novel method for high-resolution inverse synthetic aperture radar (ISAR) imaging is presented by transmitting sparse probing frequencies. In this method, only a few measurements in the range frequency and cross range time domain are needed to reconstruct the target scene by solving an inverse problem either through a linear program or a greedy pursuit. By transmitting merely a few probe frequencies instead of wideband signals, the proposed method can obtain an unambiguous ISAR image with super-resolution. The validity of the proposed approach was also confirmed using numerical simulations and real data.6. Imaging of ship targets is a hotspot in the technology research of inverse synthetic aperture radar (ISAR). The complexity of the ship motion characteristics and the difficulty caused by using the stepped chirp signal are addressed. Firstly, the digital signal processing board based on digital signal processors is designed. Then the suitable algorithm to the real-time data processing for ISAR imaging of ship targets using stepped chirp signal is selected. Finally, the task allocation and programming of the real-time signal processing on the digital signal processing board are completed. Experiment result confirms the validity of the proposed approach.