Image Processing System Based On FPGA

Digital image processing technology have a wide range of applications in daily life, production, construction, national defense and security. Traditional image processing techniques is difficult to meet the treatment effect and real-time requirements when dealing with large amounts of data. With the rapid development of chip technology, making the hardware to achieve real-time digital image processing possible, FPGA is an ideal choice for real-time image processing because of its parallel computing characteristics.This paper outlines the current situation of digital image processing, FPGA's structural features, working principle, development processes and tools. After depth analysis of digital video standards, i discuss the development process and key technologies of FPGA-based real-time image processing system. System realizes the function of video capture, image processing and image display, including video capture module, configuration module based on I~2C bus, asynchronous FIFO module, decoding module, DDR controller module, image processing module and display module, all these function modules are implement in FPGA. The video signal collected by a CCD sensor and transmitted to the decoder chip TVP5150PBS, the A/D operation converted the video signal to ITU-R BT.656 format, the asynchronous FIFO buffer the video data and then send to the decoding module, this module converted the color image to the gray image. The DDR SDRAM buffer the images frame, then send image signals to the image processing module, image signal be processed and then converted to analog signal by the D/A converter chip SAA7121H, analog signal output to display.In the image processing module, make full use of parallel computing features of FPGA, improved filtering algorithm, combine with pipeline structure to improve the processing speed of the algorithm. At the same time, use Simulink and System Generator to design the algorithm modular, these tools simplifies the implementation process, greatly improved development speed. Through the simulation and testing to verify that the algorithm can effectively achieve the desired functions. The design shows that, FPGA chips can not only be used as glue logic functions, control and connect peripheral modules, but also can replace the DSP processor to achieve image processing algorithm. Use FPGA to implement image processing is a stable, effective and economical solution.