Synthetic Aperture Radar (SAR) is an important tool for the collection of high-resolution, all-weather,all-time image data. It has more information than the opticalimage, and is widely used in the fields of military inspecting, terrain mapping,resource investigating, disaster forecasting etc.Real time high resolution SAR imaging systems are typically very complex andexpensive. They must process enormous quantities of data, and required very highcapacity data storage and high processing ability. This requires us to analyze the SARimage algorithms from the view of algorithm complexity and storage character, and todesign and realize a processing system according to the algorithms.Storage structure design is the foundation of the real time high resolution SARimage system, and it is a key problem from the system design to realization. There are3 layers of storage structure, including processor's storage structure, subsystem'sstorage structure and the system's storage structure. Focusing on the SAR imagealgorithm, all layers of storage structure problems of the SAR imaging system arestudied in this paper systematically.1. processor's storage structure: The goal of processor's storage structure is tooptimize the processor's data access structure and to improve the efficiency of thekey operation of SAR image algorithm– FFT. Main contributions achieved in thislayer are as follows:(1) Proposing two kinds of FFT parallel conflict-free data access algorithms.Due to the consideration of special data relationship of FFT algorithms, thespecific FFT processor has the benefit of high costs-performance ratio. The dataaccess algorithm is very important for the processor design, but there is no way to geteffective parallel data access until now. In this paper, a radix-2 FFT's parallel...
|