Research on Basic Algorithms of Digital Image Processing and Implementation with FPGA
|School||Harbin Institute of Technology|
|Course||Instrument Science and Technology|
|Keywords||FPGA Digital image processing Real-time|
With the rapid development of multimedia and network technology, digital image processing is widely applied in various fields, as a result, the demands on image processing are increasing. Therefore, the processing only relying on software is unable to meet the requirements, such as real-time, high efficiency and so on. On the other hand, the high speed development of microelectronics technology shows more advantages of hardware-based image processing. This paper does useful research on the methods of image processing based on FPGA, and has a positive signification for both the development of real-time image processing system and the hardware design of new image processing algorithms in the future, in which FPGA acts as the core.Firstly, the relative knowledge and basic algorithms of digital image processing are introduced. Then the logic design of some basic algorithms, such as histogram equalization, median filter, convolution, edge detection and wavelet transform are implemented using the Verilog hardware description language. The fast algorithm chosen in median filter and the tree shift-add multiplier used in convolution significantly save the hardware resources and increase the processing speed. At first, each algorithm is implemented in MATLAB, and its simulation result is used as reference. Then the logic design and simulation are achieved in Quartus II. The comparison of the simulation results in the two platforms shows that, for the same image, the simulation result of Quartus II is almost consistent with that in MATLAB.Finally, in order to carry out the logic in hardware, image processing target board is designed. The logic of each algorithm is validated in hardware and its performance is analyzed. The results shows that the FPGA-based image processing can access well processing effect, as well as its speed is far higher than the software, which can meet the requirements of real-time image processing.