Study On The Imaging Algorithm Of The Bistatic Synthetic Aperture Radar

Bistatic Synthetic Aperture Radar (BISAR) is a new type of radar, where the transmitter and receiver are carried on two different platforms. Due to the separation of the transmitter and the receiver, the BISAR can configure the geometry of platforms according to the surveillance requirements. In comparison with the conventional monostatic systems, the increased flexibility of the bistatic systems brings several additional benefits like forward-looking SAR imaging, frequent monitoring, reduced vulnerability of military applications, and ability to use multilevel interferometry, etc. In addition, the BISAR is a fundamental component of the spaceborne constellation SAR systems. Therefore, the BISAR system is in possession of a wide potential of application.Due to the separation of the transmitter and the receiver, many technical problems such as synchronization of frequency, involved adjustment of transmitter pulse versus receiver gate timing, antenna pointing, and theoretical modeling are not sufficiently solved. Besides, the imaging algorithm is another difficulty, and the purpose of this dissertation is to propose a set of precise efficient algorithms for focusing the azimuth-invariant BISAR data.The azimuth-invariant property is significant for the modern monostatic SAR imaging algorithms that the main processing steps are carried out in the frequency domain, because all point targets with the same closest range collapse to the same migration curve in the range-Doppler domain. As a result, the efficiency of these imaging algorithms is achieved by taking advantage of block processing.Because of the separation of the transmitter and the receiver, the BISAR configurations are much diverse. To rank the complexity of the geometry situation, the BISAR configurations could be divided into four cases: one fixed platform case, tandem case, translational invariant (TI) case, and general case. According to the azimuth property of the echo data, they can also be divided into the azimuth-invariant configuration and the azimuth-variant configuration. Unlike the monostatic system, the BISAR is in general azimuth-variant system as both the transmitter and the receiver move along different motion trajectories with unequal velocities. Nevertheless, the azimuth-invariant property is remained in both the TI and tandem cases.Considering the increased complexities induced by the separation of the transmitter and the receiver, this dissertation summarizes the current algorithms, studies on the problems of BISAR imaging, and proposes the following novel algorithms:First, we propose a polynomial model of the bistatic range history and approximate point target spectrum. The bistatic range history is the sum of two individual hyperbolas, which is referred to as the'flat-top hyperbola'. This makes it impossible to obtain a precise analytic point target spectrum for the BISAR data. To overcome this difficulty, the bistatic range history is approximated by a polynomial of the azimuth time. In this way, an analytic 2-D point target spectrum is derived, and efficient monostatic imaging algorithms are easily modified to handle the BISAR data.Second, a precise efficient bistatic Chirp-Scaling (CS) algorithm is proposed. Due to the range property of the BISAR data in the frequency domain, the conventional CS algorithm is hardly applied to process the BISAR data. Based on the analytic spectrum, a bistatic CS algorithm is proposed by two modified key operations: the range cell migration correction and azimuth compression implementation, which is able to achieve the high precision imaging with negligible approximations.Third, an Extended Nonlinear Chirp-Scaling (ENCS) algorithm for large baseline BISAR data is proposed. In comparison with the monostatic echo data, the range property of the BISAR data is much more intricate. The 2-D large baseline will severely deteriorate the performance of imaging algorithm under a wide swath configuration. To eliminate the affect of these factors, we propose the ENCS algorithm which greatly enhances the performance of the BISAR processing, particularly under a large baseline and a wide range swath.Finally, a sub-image auto-combination algorithm for constellation SAR system is proposed. The constellation SAR system is an important application of the BISAR technique. To enhance the image resolution of the constellation SAR system, we propose an auto-combination algorithm which is based on the image quality estimation functions.