Low-dose CT imaging technology plays an important role in clinical practice by reducing the radiation hazard to patients by reducing the scanning dose. However, low-dose CT image often contain noise and artifacts, which affect doctors' clinical diagnosis. In this paper, the gray absolute correlation degree and low-rank constraint are introduced into WESNR. The nonlocal regularization denoising algorithm first uses the gray absolute correlation degree to separate the noise points from the edge points and points in the smooth area, and filters the noise points. Contaminated blocks are coded to remove both impulse noise and Gaussian noise from the mixed noise with appropriate regularization. The experimental results show that the algorithm proposed in this paper is better than the contrast algorithm in terms of detail and edge preservation and noise suppression of low-dose CT image.