Study On Key Technologies Of Inverse Synthetic Aperture Radar Imaging

Inverse synthetic aperture radar (ISAR) is a kind of high resolution imaging radar different from the traditional radar. It has the capability of getting fine resolution images of noncooperative targets day and night in all weather from long range, so it is with great military and civil values. The key technologies of ISAR involve motion compensation, imaging algorithms, ISAR total design and so on. At present, there are still some issues of these technologies that have not been solved well and should be improved. In order to make full use of ISAR ability of high resolution imaging, and enable ISAR to accommodate with practical requirements and constantly varying application situations, ISAR technology should be deeply studied continually. In this thesis, some key technologies of ISAR signal processing are mainly discussed, including range alignment and phase compensation of ISAR motion compensation, scaling of ISAR images, and ship target imaging, which is a difficulty in ISAR imaging.Chapter 1 is the introduction. Firstly, ISAR is briefly introduced. Then, the basic concepts of ISAR imaging are elaborated and the development of ISAR signal processing technology is reviewed. Lastly, the research background and main contents of this thesis are presented.The range alignment of ISAR motion compensation is studied in chapter 2. A new range alignment algorithm that utilizes the modified kurtosis as the alignment measure is proposed, in which the amount of range shift between two echoes is estimated by maximizing the modified kurtosis of their sum envelope. Because the modified kurtosis emphasizes both the sharpness of the envelope's waveform and the spread of the envelope's energy, it is able to reflect the alignment situation accurately. Especially, this new algorithm behaves very well when target echoes have one dominant scattering center. Another new range alignment algorithm is also put forward in this chapter. This algorithm constructs the combined envelope by selecting the larger elements in each range bin of the two unaligned echoes, and a new principle based on minimum sum is utilized to measure the alignment. For target echoes with different scattering characters, the range alignment algorithm based on minimum sum all behaves well with high efficiency. The results of real data processing fully demonstrate the validity and effectiveness of the two proposed range alignment algorithms.The ISAR phase compensation algorithm is studied in chapter 3. Firstly, the method that implements the phase compensation by using the rank one phase estimation (ROPE) algorithm is presented in detail. It can obtain satisfying compensation effect and has good real time performance. Then, the phase compensation algorithm using ROPE is improved on.The blindness of setting zero as the initial Doppler frequency in ROPE algorithm is eliminated by circular shifting of the azimuth dimension in the image domain, and the precision of the translational phase estimation is increased by introducing the iteration operation into ROPE algorithm. The two measures effectively ameliorate the performance of ROPE algorithm as it directly applied to ISAR phase compensation. Compared to PGA, the improved phase compensation algorithm using ROPE requires neither a relatively good target image at the start, nor the windowing operation during the iteration, so it is with higher efficiency. And it is able to compensate high frequency phase errors as well as random phase errors.Azimuth scaling of the ISAR image is a difficulty in ISAR technology. In chapter 4, a new scaling method completely based on the obtained image for the target top-view image is proposed, which doesn't require tracking data, the prior information about the scatterers'positions of the target and so on. Generally the imaging duration is known, so the key for solving the scaling problem is to estimate the equivalent rotation rate. In this chapter, the maximum likelihood estimate of the rotation rate is presented, which is only related to the feature lines'slopes of the target shown in the image except for the wavelength. Further, by directly utilizing the radon transform to extract the slopes, the rotation rate estimation is completely based on the unscaled ISAR image, and therefore is more practicable in practice. Many real applications have demonstrated that the estimation method has high precision. By making use of the rotation rate, the azimuth resolution of the target image can be obtained as well, so is the looking angle of radar during the imaging, by which the main feature parameters of the target shape can be determined accordingly. They are very helpful to the subsequent target identification.In military affairs, the ship target is often with great values. Because the ship's motion is complicated, the ship imaging is with relatively high difficulty, where selecting the optimal imaging time is a hot topic and a difficulty all along. In chapter 5, the problem of how to select the optimal ship imaging time is discussed deeply. A method based on the estimation of rotation vectors is proposed. Firstly, the equivalent changes of the effective resultant rotation vector and its vertical component with time are estimated by imaging the data in successive partially overlapped intervals and subsequently estimating the Doppler spread and the centerline slope of the ship target shown in each subimage. Then, according to the two variables, the sea state and the rock intensity of the ship in the real environment are evaluated, and the imaging is separated into several processing cases. In each case the optimal imaging instant is determined by the two variables and the optimal imaging duration is searched out by minimzing the image entropy. The simulated data and real data verify the feasibility, effectiveness and practicability of the time selecting method, and demonstrate that high-quality ship images can be achieved under unknown sea state via this method. Especially, the top-view or side-view ship image can be obtained, which benefit the subsequent identification of ship target. In chapter 6, the ship imaging is studied by using airborne SAR real data. Effective ISAR processing methods applicable to the ship imaging are presented. Five range alignment methods for solving the slow range shift of the ship's SAR echoes are provided, including the method aligning according to the first echo, the method using the half-window accumulation, the method based on Radon transform and its simplification, and the global alignment method based on the minimum entropy. After imaging, the space variant apodization (SVA) algorithm is also introduced to suppress the sidelobes in the ISAR image. The better image quality is achieved. By processing each ship's echoes extracted from the SAR data via the presented methods, the obtained ship images are all much clearer than those shown in SAR image. This study enhances the airborne radar's ability of ship imaging.The work of the whole thesis is concluded in chapter 7 and the issues to be further studied are point out as well.