Research On Image Noise Reduction Algorithm Based On Principal Component Analysis And Bilateral Filtering

Information transmitted in the form of digital image is a major method of communication.An image is often corrupted by all kinds of noises in its acquisition and transmission,which degrades the quality of the image and impacts the subsequent image processing.Therefore,it is necessary to denoise before analyzing of image information.The main aim of a denoising algorithm is to reduce the noise level,while preserving the important signal and image features.The fusion of principal component analysis and bilateral filtering algorithm is studied in this paper.The results show that the proposed algorithm can better remove the additive noise while retaining as much as possible the edges and important signal features.The denoising method proposed in this paper is divided into two stages.In the first stage,the target pixel and its nearest neighbors are modeled as a vector variable,and the training samples are selected from the local pixel blocks by using the block matching algorithm.The principal component analysis(PCA)transform is used to reduce the dimensionality and decorrelate the redundancy of original data,and then separate the signal from the noise by threshold shrinkage function.Finally the reconstruction image is obtained by using inverse PCA transform.In the second stage,we estimate the residual noise variance of the reconstructed image to achieve the gray standard deviation for adaptive bilateral filtering.In this paper,the spatial domain denoising method and the transform domain denoising method are combined to get better denoising effect.The block matching algorithm is used to select similar local block samples,making the covariance calculated by the principal component analysis is more accurate,and then distinguish the signal from the noise over PCA domain.At the rule of optimal setting of filtering parameters,adaptive bilateral filtering algorithm based on residual noise level estimation exploits spatial parameters and gray parameters to filter a noisy image,and obtains the restored image by using local weighted averaging methods while preserving the image edge details effectively.The validity of the proposed method is verified by experiments.The BM3 D and LPG-PCA methods are used for comparison to demonstrate the effect of our method.The experimental results shows that the proposed algorithm obtains better results using objective criteria peak signal-to-noise ratio(PSNR)and structural similarity(SSIM),and the visual effect will be much improved.It indicates that the proposed method can not only smooth the noise,but also effectively preserve the details of an image.