<p>Image denoising is a widely concerned problem, which has been successfully applied in remote sensing, medicine and other fields. A typical idea of image denoising is to exploit some prior information existing in real-world data, such as low-rank prior and local smoothness prior. Some researchers devote themselves to combining both priors, however, the current methods cannot capture both properties simultaneously and adequately. Motivated by a new regularizer named three-dimensional correlated total variation (3DCTV) for robust principal component analysis problem, in this paper, we propose a new image denoising model via the double-weighted correlated total variation regularization. Specifically, we perform weighting operations on the 3DCTV regularization term and the sparse term separately, which can make fuller use of the low-rank prior, the local smoothness prior and the sparse prior of images. In addition, we add the Frobenius norm term to this model for modeling strong Gaussian noise in some real-world scenarios. Then, we develop an efficient algorithm to solve the resulting optimization problem by using the well-known alternating direction method of multipliers. Finally, we conduct extensive experiments on hyperspectral images, multispectral images and medical images under various noise situations, and the experimental results show that the proposed method outperforms the existing state-of-the-art denoising methods. Especially when the test image is polluted by low-intensity sparse noise, the MPSNR index of our method is about 5 points higher than that of the 3DCTV method.</p>