Image reconstruction for denoising based on compressive sensing

Due to the disadvantage of large amounts of data computation and image quality degradation of classical reconstruction algorithm, a novel adaptive method of image reconstruction denoising based on compressive sensing is proposed. Firstly, the wavelet approximate coefficients and detail coefficients from the image noise are Gaussian distribution, and have different variances in different levels. Secondly, the noise image is divided into image blocks of a certain size, a new compression sensing block reconstruction method has been used to recover small block coefficients. Finally, the reconstructed denoising images are obtained based on recovered detail coefficients and approximate coefficients by the separation of small block wavelet inversed transform. Experimental results show that this method is feasible and available, compared with pure wavelet denoising and block image, signal-to-noise ratio has been improved highly, the image noise has been removed effectively and the reconstructed image quality has been improved highly.