Sparse Denoising Algorithm For Video Images Based On Wireless Multimedia Sensor Networks

Wireless Multimedia Sensor Network(WMSN)is a new means of acquiring and processing multimedia information.It is widely used in traffic monitoring,target tracking and image denoising because of its advantages of low power,high scalability,diversity of perceptual information and realtime monitoring.The monitoring scenarios of wireless multimedia sensor networks are complex and changeable.Under the influence of multiple random interferences,the quality of images will decrease or even be blurred.A video and image denoising algorithm based on K-SVD and low SNR is proposed in this paper.First,the video image of wireless multimedia sensor network is preprocessed.Secondly,the key of the image is sparse representation by the redundant dictionary of DCT,and the residual frame of the image is sparse de-noising through the redundant dictionary of DCT.Finally,the key frame is reconstructed by the orthogonal matching tracking algorithm based on the Dice criterion,and the key frame and the residual frame after the denoising are superimposed to realize the video image denoising of the low signal to noise ratio wireless multimedia sensor network.Experimental results show that the algorithm can effectively denoise and achieve better visual effects.However,the above methods do not make full use of the structural similarity between image blocks,which affects the effect of image denoising to a certain extent.In this paper,a video image denoising algorithm based on image clustering and non local regularization is proposed.This method makes full use of the sparsity of the image and the non local self similarity to construct the sparse regularization model,and the wireless multimedia sensor is used as a substitute function of the iterative weighted contraction.The network video image is reconstructed by denoising.The experimental results show that this method can remove a lot of noise,and can better retain the details of the structure of the image.It has better visual effect and is more practical in a complex environment.