The Design Of C6000-Based Image Processor

In this Dissertation, the digital image processing technology is explained in detail, and the working principle and structural characteristic of an image processor are described. A systematic design of the real-time image processor based on C6000 is presented according to the requirements of practical project. The performance of the real-time image processor was proved by the results of practical tests.The high-speed digital signal processor TMS320C6202 is a kernel device in this system. The TMS320C6202 has a performance capability of up to 2000 million instructions per second (MIPS). It satisfies the needs of real-time image processing even if a complicated arithmetic is applied, and so the effect of target recognition and target tracking is improved. This DSP can transmit and deal with data in parallel by DMA without CPU, so processing speed and efficiency is advanced.The use of a video A/D converter, SAA7114H, makes maximum convenience for application, including line-locked clock generation, anti-alias filter and comb filter.A real-time median filter is achieved by FPGA, which can save time to let DSP deal with more complex arithmetic. SDRAM is used as image storage, which can store many frames image data to provide the foundation for the correlation operation between frames. FIFO is a data buffer for real-time image. All of these make the image processing system simple design, decreasing the write-read time and solving the question of transmission and delay, and allowing the data processing and the data grabbing at the same time. The processed results can be output in 20ms, so the bandwidth of the C6000-based image processing system is improved.The C6000-based image processor is used in different kinds of tracking system and scenes. A lot of debugging work has been done on hardware and software, and finally the practical tests have been done in an airfield. The results of the tests showed that the C6000-based image processor can satisfy the requirements of real-time target tracking.