The Real-time Image Processing Embedded System Based On The ARM、FPGN、Multi DSP

With the rapid development of modern information technology, the research andapplications in image processing had be received intensive attention, especially in thefield of real time image processing. In recent years, the embedded system and DSPtechnology had made great progress, which help the research outcomes of digitalsignaling processing been implemented in practical systems. Therefore, the newtheory is obtained, which is every important for the development in the imageprocessing area. However, the practical embedded systems require stronger computationability, higher data transfer rate throughput and mulit task real time processing in imageprocessing, Radar signal processing, radio software, electronic countermeasures and3Gsimulation calculation. Therefore, when a signal DSP cannot satisfy the requirement ofreal time and computation speed, multi DSPs may be introduced to be cooperatedtogether. In this thesis, an embedded system combining ARM, FPGA and multi DSPs isdesigned and it is worthwhile both in theory and in practice．This paper mainly focuses on the following aspects：In the first part, the thesis introduced the latest development in the real time imageprocessing area. Comparing the state of art research outcomes in this area, theadvantage and the disadvantage by using ARM or DSP as real time image processingkernel respectively are obtained. We design a general real time image processing systembase on ARM, FPGA and multi DSPs to fulfill the features of real time, high speedcomputing, high data through put and re constructable in our system. The system can beused in the high speed image processing and recognition area by adding or cutting.In the second part, a framework of the system is introduced and then the relatedtechnologies are illustrated. The technologies include the architecture of Nios CPU, theAvalon bus standard,the design method based on Nios Ⅱ multi CPU.In the third part, the hardware design and implementation are explained. These are,a FPGA end video input/output module, a DSP storage module, an interface modulebetween FPGA and DSP, a power supply module. Next, the features of the embeddedLinux and the driver coding in Linux system are presented. And then, theimplementation of system DMA communication and GPIO communication driver is provided.In the fourth part, two applications based on the designed system are brieflydescribed. They are ultrasound image processing applications and video surveillanceapplications. Finally, the achievements and shortcomings of this research aresummarized. We also had given a new future application.