| Synthetic aperture radar (SAR) is used to form fine-resolution images that are maps of radar reflectivity of some scene of interest, from range soundings taken over some spatial aperture. It can provide some specific information in all day and all weather which optical and infrared sensors can't provide. Therefore, it becomes an important remote sensing tool and is widely used in both the military and the civil application. High resolution is the most important feature which makes SAR distinguished from traditional radar, and higher resolution is the unremitting pursuit in SAR development. With the development of sophisticated SAR signal processing and imaging methods, more specialized radar problems are being studied in the framework of SAR system. A prominent example is ground moving target detection and imaging, which has become a hot topic in recent years.Image formation algorithm is the core of SAR signal processing. A number of algorithms have been developed to form fine resolution images of large scenes. Among them, polar format algorithm (PFA) is polular because of its high imaging efficiency, efficient nonplanar motion compensation capability, and moving target's linear range walk correction capability. Unfortunately, the classic PFA, due to wavefront curvature, has limits on the scene size and geometric fidelity. These limits can be alleviated by wavefront curvature compensation technologies, such as subaperture processing and spatially variant post-processing. However, all the proposed methodologies adopt approximation in the wavefront curvature compensation. Therefore, the modified PFA still has scene size limits that are especially severe in ultrahigh resolution imaging or mosaic imaging (such as image mosaic in stripmap SAR and circular SAR). In this dissertation, the modified PFA is further improved and extended to solve the high resolution imaging problems when SAR is under different situation, such as on different platform, in different modes or with complicated motions.Chapter 1 is the introduction. Firstly, the history of SAR technology is outlined. Then, the latest developments of ultra-high resolution SAR and SAR/GMTI in China and abroad are reviewed. Lastly, the research background and main contents of this thesis are presented.In chapter 2, the polar format algorithm (PFA) is studied. Firstly, the principle of PFA is briefly introduced. Then, the range cell migration correction process in polar format transform is studied, which shows that the polar format transform can also be interpreted as a combination of a range scaling and an azimuth scaling. Especially, the azimuth scaling includes a keystone transform, which is originally used to correct the unknown linear range migration of moving target. Based on the scaling transform interpretation, a novel implementation of PFA using the principle of chirp scaling is investigated. Finally, since the differential range was approximated during the analysis about the wavefront curvature in the traditional PFA, it could not meet the precision requirement. To this problem, a more accurate derivation on the wavefront curvature is presented, the detailed wavefront curvature compensation technologies will be discussed in chapter 3.There are two kinds of technologies about the wavefront curvature compensation in the present literatures, one is based on the subaperture processing, and the other is based on the spatially variant post processing. In chapter 3, both of the algorithms are improved, and the modified versions are proposed, which are based on the wavefront curvature analysis presented in chapter 2, are proposed. Finally, the two new algorithms are compared with each other in terms of the accuracy of wavefront curvature compensation, computer complexity, and required precision for inertial navigation system.In chapter 4, the modified PFA with spatially variant post processing is applied to several SAR modes, including ultra-high resolution spotlight SAR, stripmap SAR, and circular-scan SAR. Firstly, a new autofocus algorithm, which can work in ultra-high resolution SAR mode, is proposed. Combining this autofocus algorithm with the modified image formation algorithm, the signal processing flow for ultra-high resolution spotlight SAR mode is presented. Then, by extending the spotlight SAR image formation, an imaging processing method based on automatic, seamless mosaicing of neighboring subpatches is presented to generate the high resolution full-strip image with high geometric accuracy in the stripmap SAR mode. Finally, the image formation in circular-scan SAR mode with large maneuvering of radar platform is studied. By extending the wavefront curvature analysis to the diving trajectory case, a modified PFA with wavefront curvature compensation is presented. Then, using the modified PFA, a circular-scan SAR image formation algorithm, which can even work with large maneuvering of radar platform, is proposed.In chapter 5, the SAR signatures of moving target are presented. Firstly, the SAR image-error-spectrum model of the moving target is defined. Then, the image-error-spectrum after the RMA and the PFA processing, respectively, are derived. Based on the formulation of image-error-spectrum, detailed analysis on the SAR signatures of moving target, including the geometric displacement, residual range migration, and the defocusing effect in both range and azimuth dimensions is performed. Finally, the simulation results validate the theoretical analysis.In chapter 6, some key problems in multi-channel SAR/GMTI system are studied. In chapter 5, it has been shown that PFA not only can produce high resolution SAR image of stationary scene, but also can eliminate all the linear range walk of moving target at the same time. Therefore, a three-channel SAR/GMTI signal processing method based on PFA, which can solve the moving target's range walk problem in high resolution SAR/GMTI system, is proposed. Furthermore, by adjusting the parameters of polar format interpolation, the phase difference of stationary target in different channels can be corrected automatically, so the traditional phase compensation process before clutter cancellation is not required in this method. Finally, the simulation results prove the validity of the method.In the end, the work of the whole thesis is concluded and the issues to be further studied are pointed out as well.
|
PDF Full Text Request