Parallel Implementation Of Airborne SAR Image Formation On Hardware

Synthetic Aperture Radar (SAR) is a microwave imaging system which is widely used in the fields of both military reconnaissance and civil remote sensing. With the development of SAR imaging techniques, high image precision and high real-time rate have becoming more and more important, especially on military field. There are two methods on real-time imaging, one is the special SAR processing system and the other is the general-purpose parallel processing system. Comparing with the first method, the second one has the advantages of shorter developing time, lower cost and easier upgrade. This paper focuses on parallel processing of SAR imaging algorithms and two parallel imaging methods based on the existing sequential algorithms are presented and implemented on a general-purpose parallel processing system. The main contents of this paper can be concluded as following:Chapter 1 is the introduction. The history of SAR technology is outlined including airborne and space-borne radar systems, and then the main SAR modes and parallel processing are also introduced.Chapter 2 discusses SAR imaging theories and analyzes the SAR azimuth resolution. Then two SAR imaging algorithms, spotlight mode SAR convolution backprojection (CBP) imaging algorithm and circular-scanning mode SAR imaging algorithm, are researched, and the images of raw data are presented.Chapter 3 introduces the fundamentals of the parallel algorithm.Then we introduce the hardware and software recourses of the TS-C43 card, and discuss the program developing steps based on the general-purpose parallel processing system.Chapter 4 mainly studies the parallel design of the SAR imaging algorithms according to the recourses of the parallel processing system introduced in chapter 3. Then we implement the parallel SAR imaging algorithms on the parallel system, and analyze the performance of the parallel algorithms. Finally SAR images obtained by the parallel algorithms from raw data are also given.Chapter 5 concludes the work of this paper, and points out future work.