Research And Implementation Of Algorithms In FPGA-based Image Processing System

With the development of computer technology and Internet, digital image processing technology has penetrated into all areas of science and technology. So that people’s lives couldn’t been separated from the image processing technology in modern society. Today digital image processing technology mostly used an early special processing chip, the processing speed and performance of it increasingly difficult to meet the development needs of the people. In data processing, the FGPA digital data processing system have the characteristics of strong real-time and strong logic resources. So it has a unique advantage in the way of the design of data processing system.In my paper, I choose Altera’s Stratix Ⅳ GX series of chips DE4 as the hardware platform of image processing system and combining the parallel characteristic and pipeline operation of the FGPA. Firstly, an introduction of the hardware architecture design of the entire real-time image acquisition and display systems has been given, it’s mainly about the configuration of camera interface, camera sequence control, data acquisition, format conversion, memory control, DVI video signal synthesis and Gray-scale synthesis. It also analyzes the working principle of DDR2 SDRAM and FIFO memory design. Secondly, it analyzes and studies the realization principle of common image preprocessing algorithms. In the removing noise aspect, it proposes Fast Switching Median Filter Algorithm which based on an introduction about the traditional median filtering bubble method and General fast median filtering algorithm. In the image sharpening aspect, it introduces the most popular sharpening algorithm, differential method, and then Laplace sharpening algorithm and Laplace algorithm hardware with the template center 5 has been proposed. Finally, in order to complete debugging, the pre-processing algorithm module is embedded into the entire real-time display system.We have finished some exploratory works related to real-time image processing on FPGA areas, which has a positive reference value in this field.