Study On Imaging Methods For SAR On Curved-Path Platforms

With the rapid development of the radar technique, the research on the curved-path synthetic aperture radar (SAR) has a special application value. Compared with the motion errors of the traditional SAR with uniform linear motion, the curved-path SAR platform has a large spatial three-dimensional (3-D) acceleration to realize the flexible control and multiple applications. With the trend of curved-path SAR system, we start our study with the generalized SAR signal model. The main content of this dissertation is summarized as follows.1. The imaging model and properties are studied because the traditional ones are invalid for curved-path SAR. Under the condition of curved-path, radar cannot fly along a straight line with a constant velocity; thus the imaging model and properties based on the traditional SAR have great degradations for curved-path SAR. To solve these problems, a geometrical model of the curved-path SAR with general configurations is graphically illustrated by vector notation; Based on this accurate model, the mathematical descriptions of the steering commands are derived. Then, by using the gradient method, we study the intrinsic curved-path SAR properties, including the derivation of the accuracy expressions of the bandwidth and resolution, system constraint quantitative analysis, and the effects of the model mismatch on the imaging. The analyses in this part are the bases for the algorithm design and provide tools for the SAR system design. Particularly, applications of the model and properties for different modes are discussed and corresponding analyses and descriptions are given.2. A modified wavenumber domain algorithm, based on equivalent range history, is proposed for the highly squinted spotlight SAR with curved-path to improve the quality of the focusing image and realize the high squint and high resolution SAR processing. We firstly analyze the signal model of the spotlight SAR with curved-path by using time-frequency analysis, and list the variations of the inner structure and the problems, i.e., spectrum aliasing and traditional wavenumber domain algorithm fails in the curved-path cases. To remove the spectrum aliasing, by using the new time-frequency rotation functions which are constructed to adjust the sampling interval, the preprocessing method is proposed based on the traditional two-step technique. According to the property of the wavenumber domain algorithm, i. e., the linearized phase, equivalent range history is derived with a form similar to that of traditional SAR, and it can avoid the use of the method of series reversion (MSR) to obtain the two-dimensional (2-D) spectrum and the linearized phase. With the incorporation of the preprocessing, a modified wavenumber domain algorithm is presented for curved-path spotlight SAR.3. The frequency domain algorithm (FDA) and imaging algorithm based on 2-D Keystone transform are respectively proposed for the airborne SAR and spaceborne SAR with several problems such as spectrum aliasing and spatial-variations. The generalized imaging methods were studied for the curved-path beam steering SAR. With the flexible beam steering commands, the stripmap, spotlight, sliding spotlight, and Terrain Observation by Progressive Scans (TOPS) SAR modes emerge as the times require, and we call them as beam steering SAR for short. The geometrical model is firstly built to analyze the impact factors for curved-path SAR, and a FDA, based on updating processing, is proposed with incorporations of azimuth pre- and postprocessing to solve the problems such as spectrum aliasing, time domain aliasing, and spatial-variations. It can greatly improve imaging qualities. Different from the airborne SAR, the’stop-go’assumption should be considered in the spaceborne SAR. An accurate mathematical description is derived with considering the’stop-go’assumption and imaging methods based on 2-D Keystone transform are presented. The key of these methods are the derivation of the linear relationships between the zero-/first-order and higher-order terms. These methods can be well suitable for all the orbits with different heights and have advantages that the procedures are simple and clear. Particularly, the processing methods for beam steering SAR have widely applicability, completeness, and expansibility.4. In order to solve the problems of spectrum aliasing and spatial-variations in the general bistatic SAR with curved-path, the generalized polar format algorithm (GPFA) and the extended azimuth nonlinear chirp scaling method are proposed. In the GPFA, The pre-and post-processing operations for monostatic SAR are extended to be suitable for the bistatic cases and the interpolation function is incorporated to focusing bistatic SAR data. It can well solve the problems such as spectrum aliasing, time domain aliasing, and spatial-variations and is efficient and easy to implement for curved-path bistatic SAR. An EANCS is proposed, which is an extension of the traditional ANCS that cannot be applied for the several SAR modes. According to the nonlinear relationship between the Doppler centroid and the Doppler paramters of the Taylor series expansion, one can obtain the chirp scaling factor which cannot be derived in the traditional ANCS for the general bistatic SAR. With the incoporations of the polyphase filter and perturbation function into the EANCS, the degrees of freedom increase and the spatial-variation can be eliminated entiredly. It can greatly improve the imaging qualities and range for the general bistatic SAR with curved-path.