Research On Spaceborne SAR Fast Imaging On Multi-core Processor

In recent years,synthetic aperture radar(SAR)is moving towards high resolution and wide mapping belt,which makes a huge surge in the satellite remote sensing data.This development also has a higher requirement for the integrated processing performance of the processing platform.On the other hand,portable software based remote sensing data processing system has become an important development trend in the radar processing system at home and abroad.How to construct a high performance and low power remote sensing data processing system which is easy to be transplanted has become an urgent task to solve.According to the above problems,this paper mainly has done the following aspects:(1)Since the current construction of remote sensing data processing platform processing has large amount of data and mass storage,the algorithm is complex and flexible.The high requirements of real-time processing and the application environment is complicated and the system is general good.The paper analyzes and compares the advantages and disadvantages about different chips in their processing capacity,storage capacity,power consumption,environmental adaptability,IO capabilities and other capabilities.The paper also analyzes the characteristics of the difficulty of system development,application and expansion and versatility and selects the appropriate technical approach for the building of the Space borne SAR remote sensing data processing system.(2)This paper chooses the way of "Power PC embedded processor +Linux embedded system" to build the remote sensing data processing system.Main work includes:(1)Study on multi-core processor hardware,build the embedded remote sensing data processing system and build the cross compile environment;(2)researching the development and configuration of application layer,method of multi-core and multi-threaded software operation,different layers of software and so on;(3)verification on the stability of the system and the rate of hardware accelerator.(3)The paper take space borne SAR imaging algorithm for CS as an example to verify the efficiency of Power PC multi-core processor platform.By analyzing the characteristics of CS algorithm and the time consuming steps,this paper studies its complexity and parallelism,and designs the optimization program of multi kernel and multi thread software.Finally,it is implemented on the multi-core processor hardware platform.The test results show that the technical approach has a good effect.