Image Denoising Based On Wavelet Transform And Its Fpga Implementation

Image processing techniques have developed to be a comprehensive subject with wide application and abundant contents. Its development bases on two research fields, which have different proprets, but compact relationship:image processing algorithms and corresponding circuit implementation.General speaking, there are 3 kinds of image processing systems to implement the digital image processing algorithms, the main chips of the each system are ASIC(Application Specific Integrated Circuit), DSP(Digital Signal Process chip) and FPGA(Field Programmable Gate Array). Comparing the other two systems, the image processing system with FPGA is especially good at pre-processing for images because of its high speed and the ability to process a great amount of data.In this paper, a image denoising processing system using FPGA has been designed. Firstly, we analyzed and studied the image denoising algorithm based on wavelet transform,then researched the image denoising using neighbouring wavelet coefficients(NeighShrink), and gived the advantanges and disadvantages of the NeighShrink by matlab simulation. In order to improve the NeighShrink algorithm,which is to use a suboptimal universal threshold and identical neighbouring window size in all wavelet subbands,in this paper two improved methods are given. The one classified the wavelet coefficients by least squares support vector machine(LS-SVM) before doing NeighShrink,and the other can determine an optimal threshold and neighbouring window size for every subband by the Stein's unbiased risk estimate (SURE) and the inter scale dependencies of the wavelet coeffieciens.The result of simulation shows that,compared with traditional method,these two methods are effectively on image denoising and can keep the edge information at the same time,especially in high noise situations can get better peak signal to noise ratio.Secondly,we simplied the improved NeighShrink algorithm to satisfactify the low noise and implemenet on FPGA. Finally,we design the FPGA-base wavelet denoisng system and discuss the implementations of function circuit in FPGA,including 2-D discrete wavelet transform unit,2-D inverse discrete wavelet transform unit,SDRAM control unit,denoising calculate unit and the sysytem control unitThe simulation results of every unit and the whole system show that the wavelet denoising sysytem given by this paper can meet the actual image processing requirments and have remarkable theoretica significance and application value.