Compressedsensing denoising algorithm for astronomical image

Jie ZHANG,Yi ZHU,Xiaoping SHI

DOI: https://doi.org/10.11918/j.issn.0367-6234.201609002

2017-01-01

Abstract:In the deep space exploration, astronomical image acquisition, transmission and processing have always been the focus of research. To solve problems of slow convergence speed, poor denoising performance in compressed sensing iterative shrinkage-thresholding algorithm for image processing, an improved iterative shrinkage-thresholding astronomical image denoising and reconstruction algorithm with high performance is proposed. Firstly, the BB linear search stepsize of the classical steepest descent algorithm is used to accelerate the convergence speed of iterative shrinkage-thresholding algorithm;secondly, to further improve the reconstructed astronomical image quality, based on the classical VisuShrink shrinkage-threshold, a decreasing VisuShrink shrinkage-threshold is proposed to select the image information; since the pseudo-gibbs effect caused by threshold denoising method will appear in the process of image reconstruction, the cycle spinning method is finally employed to adjust the reconstructed image in each iteration. Multiple experimental results show that, compared with the traditional compressed sensing iterative shrinkage-thresholding algorithm, the algorithm proposed can not only obtain better denoising performance and faster convergence speed, but also effectively protect the astronomical image detail information, such as feature and texture. In addition, when compression sampling ratio is lower, the algorithm proposed also can obtain relatively higher peak signal to noise ratio and visual quality, proving the effectiveness of the algorithm proposed for astronomical image denoising.