Research On Image Restoration Based On Lucy-Richardson Algorithm

In daily life,pictures become an important way of information transmission.With the increase of the amount of information transmitted and the increase of the number of pictures,all kinds of image blur problems are inevitable.The phenomenon of fuzzy image makes people unable to obtain effective information from it,which brings a lot of inconvenience.In order to solve the problem of image blurriness and enable the image to transmit more effective information,this paper restores the blurred image based on Lucy-Richardson algorithm,so that the image can be restored to its original state as much as possible.Aiming at the problem that the number of iterations of Lucy-Richardson algorithm cannot be determined,a method for image restoration based on image quality evaluation parameters is designed.The method firstly determines the range value of image quality evaluation parameters,and then calculates the PSNR,SSIM and STD values after each image iteration.The calculated parameter value is compared with the range value.If it is not in the range value,the number of iterations is increased once until all the three parameter values are in the range.Finally,an effective restoration image is found to reach the target range.The results show that this method can accurately and effectively screen out the restored images that meet the requirements,and the operating efficiency of the algorithm is also significantly improved.In order to solve the problem of amplification noise caused by several iterations of Lucy-Richardson algorithm,a further image restoration method based on the combination of LR algorithm and haar wavelet algorithm was designed.Firstly,the degraded image is decomposed by two-dimensional wavelet,and then the first LR restoration is performed in the low-frequency domain of the two-dimensional decomposed image.Then,the two-dimensional wavelet reconstruction is performed,LR restoration is performed again in the lowfrequency domain of the reconstructed image,and then the one-dimensional wavelet reconstruction is performed,and finally the restored image is output.The results show that this method can effectively suppress the noise amplification problem of LR algorithm and improve the clarity of image restoration.Finally,the two methods are combined to solve the problem of noise amplification in the low-frequency domain,and at the same time,image quality evaluation parameters are added to find the optimal number of iterations,so as to achieve the best image restoration effect.This method effectively avoids the problems of long running time and low running efficiency caused by too many iterations.