Research On Image Restoration Algorithm Based On Alternate Kalman Filter

People can quickly and accurately understand the information contained in a clear image.However,due to the presence of noise in the imaging system,especially the universal Gaussian noise,it is quite difficult to clear image.The degraded image caused by noise brings some difficulties for image analysis,image understanding and other post-processing.Therefore,image noise reduction technology has become particularly important in digital image processing technology.In order to reduce the noise,present the detailed edge information of the restored image clearly and shorten the running time of the algorithm,an alternate Kalman filter image restoration algorithm is proposed.The specific work is as follows:1.By analyzing the typical noise characteristics and their mathematical relations,the histogram and probability density distribution map corresponding to the typical noise pollution image are given by MATLAB.The histogram and probability density function distribution map of the image polluted by noise are analyzed and compared.The subjective evaluation criteria and objective evaluation criteria for image restoration are discussed,and the mean square error and peak signal-to-noise ratio are selected as the objective evaluation criteria for image restoration.2.The typical image restoration methods,such as Wiener restoration and Kalman filter restoration,are analyzed,the typical image restoration methods are simulated,the advantages and disadvantages of various methods in mean square error,peak signal-to-noise ratio and calculation time are compared,the relationship between Wiener restoration effect and restoration parameters is studied,and the fitting curve defines the appropriate restoration parameters.By combining Wiener restoration with Kalman filter restoration(Wiener-Kalman filter restoration method),the mean square error,peak signal-to-noise ratio(PSNR)of the restored image is equivalent to Wiener restoration,but the subjective evaluation shows that the restored image obtained by Wiener-Kalman filter restoration method is clearer.3.Aiming at the problem that the Kalman filter restoration method has the advantage of short calculation time but the restoration effect is not as good as the Wiener restoration method,an alternate Kalman filter restoration algorithm is proposed.The pixel information between adjacent rows(or columns)can not be used.The algorithm takes the first row(or column)of the matrix(image)as the initial parameter of the Kalman filter prediction equation,and then the first column(or row)of the restored image as the initial parameter of the second Kalman filter prediction equation.The simulation results show that the improved Kalman filter restoration algorithm is 1/50 of the Wiener restoration time,which significantly shortens the restoration calculation time.Compared with the Kalman filter restoration method,the alternate Kalman filter restoration method has strong noise reduction ability.