Three-Dimensional Reconstruction Technique Based On 2-Dimensional ISAR Image Sequence

ISAR (Inverse Synthetic Aperture Radar) is a high-resolution imaging radar. It's vertical resolution is achieved by transmitting broadband signals and its cross-range resolution is achieved by targets'rotation and Doppler information, obtaining two-dimensional high resolution. Inverse synthetic aperture radar can imaging ships, satellites, aircraft, missiles, and other moving targets, its 2-D imaging can not only achieve long-distance, but also all-time and all-weather. With the development of the ISAR imaging technology, improved imaging resolution is getting more and more attention, leading to the research of 3-D imaging techniques. 3-D images can provide the robust features and characteristics of the feature points on the target, making it easier to identify the target, playing a very important role in obtain moving targets'information.This thesis mainly investigates 3-demensional imaging of moving targets based on 2-demensional image sequences of single-antenna ISAR. The concept is proposed which is applying the factorization method to ISAR 3-demensional imaging area. Analysis and simulations are done. Main research contents are listed as below:(1)The basic principles of two-dimensional ISAR imaging of three-dimensional targets are introduced and the simulation experiment is done, then clear 2-dimensional plane image is showed after the range alignment and phase compensation of the real data. The information above laid the foundation for later analysis and processing.(2)Usually because ISAR target is non-cooperative, cross-range resolution need to be estimated by the cross-range scaling algorithms, achieving cross-range scaling. Three-dimensional image reconstruction makes sense after all images are processed by cross-range scaling. In this paper, the cross-range scaling algorithm based on fractional Fourier transform (FrFT) is studied. Estimate the target angular velocity of rotation during the imaging using FrFT to find the total rotation angle during imaging, achieving range-cross scaling of two-dimensional ISAR images.(3)ISAR image is not as clear and easy to distinguish as the optical image, so it is required to choose some special strong scattering points to reconstruct targets'shape during 3-demensional reconstruction. The selection and matching of feature points are required. Aiming to ISAR image, we first select a part of strong scattering points in the image, then divide the image by same grids, and select same number of strongest scatter in each grid, which names grid method. The matching work so far is carried out manually.(4)The factorization method and sequential factorization method is studied and applied in ISAR 3-demensional imaging experiments. The simulation proved the feasibility of the two methods and the feasibility to apply both of them to ISAR 3-dementional imaging. The comparison of accuracy between the two algorithms is carried out then. At last, the factorization method is applied to real data and proved to be feasible.