Unitary ESPRIT Based Superresolution ISAR Imaging And Its Applications

Inverse synthetic aperture radar (ISAR) has the ability to obtain two-dimensional, high-resolution images in all-time and all-weather conditions which can reflect size, shape, structures and attitude of the targets, and other useful characteristics information. ISAR imagery plays an important role especially in military applications such as target identification, recognition, and classification. In these applications, a critical requirement is to achieve high resolution in both range and cross-range domains. Higher frequency bandwidth offers better resolution in the range direction; meanwhile, higher the angle width offers better resolution in the cross-range direction. For high-resolution ISAR imaging problems, this paper discusses the subspace-based estimation of signal parameters via rotational invariance techniques (ESPRIT) algorithm detailed. With respect to the previous work, this paper focuses on the following aspects:Super-resolution imaging is a kind of technique used in radar for obtaining much super resolution images even with a small signal bandwidth and small range of observation angle, which is usually achieved with large bandwidth and large observation angle range. For ISAR imaging, the higher the resolution of the image is, the higher the rate of follow-up target detection and identification can be obtained. Firstly, based on analysis of the imaging principle and establishment of ISAR imaging model, translational motion compensation is done with the help of classical techniques. Method of equivalent edge currents (MEC) is used to calculate the scattering of four aircrafts. With adoption of the Range Doppler (RD) algorithm, favorable ISAR images of the aircrafts are obtained.Unitary ESPRIT based super resolution ISAR imaging algorithm is emphatically discussed. The main idea is to transform the complex-valued ISAR data into real-valued data. The thesis analyzes the performance of the unitary ESPRIT, MUSIC and conventional ESPRIT methods statistically. And also the properties and influence of ID ESPRIT, 1D unitary ESPRIT,2D ESPRIT and 2D unitary ESPRIT are discussed when they are applied into radar imaging. Furthermore, unitary ESPRIT algorithm processes the matrix operations in real number field which is much less complicated than that in complex fields, and easy in realization of projects.A novel method employing 2D unitary ESPRIT and spatial smoothing technique is proposed for super resolution ISAR imaging. By utilizing the spatial smoothing technique, this contradiction that 2D unitary ESPRIT requests multiple snapshots while ISAR echo data only has one snapshot, is resolved.Since the motion parameters of non-cooperative target are usually unknown, it may result in difficulties in cross-range scaling of ISAR images. Thus, two different cross-range scaling methods are proposed:Cross-range scaling based on rotary match and Cross-range scaling based on chirp rate estimation. The former first employ JTF technique to get instantaneous ISAR images of target at several different time, then 2D unitary ESPRIT method is utilized to extract scattering centers from the ISAR images obtained. Searching the angular velocity via rotating the scattering centers extracted from one ISAR image to another, the right estimator of angular velocity is obtained only when the two sets of scattering centers match each other best, thus cross-range scaling factor is obtained. The latter consider that each scattering center in the ISAR image carries rotating information of the target via the chirp rate, therefore combining 2D unitary ESPRIT scattering center extraction method with FrFT technique together to estimate the chirp rate and angular velocity, further for cross-range scaling. Thanks to the super resolution performance of the proposed 2D unitary ESPRIT scattering center extraction method, high accuracy of cross-range scaling can be guaranteed.For classification and recognition, a new noise-robust method for the radar HRRP reconstruction via unitary ESPRIT is presented. This method is accomplished by first using unitary ESPRIT method to extract the scattering centers accurately. Then, reconstruct the HRRP by means of the extracted scattering centers, i.e. locations and amplitudes. Experimental results show that the proposed classification scheme has better robustness, recognition rate and anti-noise performance than the original method which based on IFFT.