Abstract: This paper introduced the local pixel grouping princical component analysis to video denoising and kept the correlation of the video sequence to suppress the noise of the video by 3D block-matching wavelet thresholding algorithm.The image of the first stage of the local pixel grouping princical component analysis is replaced by the result from wavelet thresholding of 3D block-matching video denoising algorithm to avoid the local effect of local pixel grouping princical component analysis for the video.Finally,local pixel grouping princical component analysis also suppresses the block effect of wavelet thresholding of 3D block-matching video denoising algorithm.Experimental results show that our algorithm introduces the princical component analysis to video denoising and solves the block effect of 3D block-mat-ching video denoising algorithm better.The subjective and objective comparison between different algorithms also proves that our algorithm has better denoising effect.

Key words: Video denoise,Princical component analysis,Block-matching,Wavelet thresholding