Research And Implementation Of Blind Image Restoration Technology Based On Parameter Estimation

Because images are highly intuitive and easy to understand,they become the main source of information for people to perceive and understand the world.The survey shows that more than 82% of information is obtained through human vision.However,in the imaging process,the image is degraded due to the relative displacement of the object and the image capture device,which is called "motion degradation".The clear image is the key to accurately obtain information,so it is of great significance to study the restoration of motion-blurred images.There are two main steps to complete the restoration of motion blurred images: The first is the estimation of blur parameters,including angle parameters and scale parameters.The second is to restore the image based on the completion of parameter estimation.The error of parameter estimation will determine the quality of the restored image.The main research results of this article are as follows:(1)Point spread function(PSF)parameter estimation for motion blurred images.This paper proposes an improved PSF estimation algorithm based on Radon transform for images with more tones and single regions;through the motion blurred image Perform block processing and use edge detection to find sub-blocks with more contour information.After the PSF parameter estimation is performed on these characteristic image blocks separately,the two sets of data with the smallest gap are selected to calculate the arithmetic mean value as the estimated value of angle and scale.Eliminate the greater influence of a single area of tone on the true value.It also uses morphology to perform a closed operation on the frequency band map of the blurred image to eliminate the cross-shaped bright line interference,and then sets a threshold to corrode the image to varying degrees.Through the data analysis of experimental results,the algorithm in this paper improves the estimation accuracy of PSF parameters.(2)Due to the unequal length and width dimensions of most images in reality,this paper proposes to use the preprocessing-based cepstrum theory to estimate the PSF parameters of the motion blurred image.Specifically,the two-dimensional map of the blurred image spectrum is used to solve the blur angle.Figure solves the fuzzy scale,solves the obstacles of unequal length and width of the image,and verifies the feasibility of the algorithm through simulation experiments.(3)Based on the work of(1)and(2),use inverse filtering,Wiener filtering and Lucy-Richardson algorithm to restore noise-free images and noisy images,and restore the restoration through objective evaluation of image quality The latter images are compared and analyzed in detail,and it is verified through simulation that the parameter estimation algorithm proposed in this paper is effective for blind image restoration.