The Parallel FFT Algorithm FPGA-based And Deblurring Research

With the widely spread of the camera and the development of digital camera function,the image processing technology has developed rapidly.Applications will become more and more wide,almost covers the astronomy,geography,medicine,and so on various aspects of life.In the processing of taking photo,because of some accidental factors often cause the image of the fuzziness,led to the image covered by the unuseful information.Image deblu is to help people to solve the problem of image fuzzy technology,because of the huge demand,digital image processing become one of the hottest topics in recent years.The general research method of image deblurring is the spatial domain method and the frequency domain method.In which,the frequency domain method is convert the spatial domain image to the frequency domain,then process blur image in the frequency domain.Now,it is a research hotspot in the field of image deblurring.Simple motion blurred image can keep the linear shift invariant system basic in the image processing,so it usually has important image information in the frequency domain.Therefore,this article select the motion blurred image to do the frequency domain processing research,and put forward a method of estimating the point spread function(PSF)based on the kind of the binary motion blurred image spectrum.The experiment results show that using the estimate the PSF combine with the classical image restoration method can get high quality image restoration.Because fuzzy image frequency domain transformation based on FFT algorithm is very important in the process of image deblur algorithm,many critical applications in image processing usually use the frequency domain transform processing based on FFT.In this paper,We Select studied the parallel FFT algorithm implementation method based on FPGA,in which,based on the parallel characteristic of hardware system,using Xilinx FPGA IP core,realizing the high efficiency FFT algorithm based on FPGA,implementing the FFT parallel processing based on the streaming structure,greatly increased the speed of FFT algorithm,reducing the computation time required,achieving the goal of increasing the FFT computation speed.Using Matlab and ModelSim to simulate,the results show that the new scheme can improve the operation speed in the condition of ensure calculation accuracy and implementation complexity.The study has some value for the realization of FFT hardware algorithm and real-time image processing based on FPGA in the future.