The invention discloses a gradient-related TV factor image denoising and deblurring method based on a noise level. An image obtained by daily shooting exhibits degradation of a certain degree; even ifa static target is stably imaged, the obtained image also comprises a pixel-level fuzzy quantity, and in addition, noise is inevitable; and a blurred image with noise can be subjected to a reset operation to improve image quality, however, noise inhibition and deblurring have a contradiction, and the balance of the noise inhibition and the deblurring needs to be given a consideration during imagerestoration. The method puts forward an image denoising and deblurring method based on total variation regularization, while the image is subjected to fuzzy recover, the characteristics of a Richardson-Lucy algorithm are combined to obtain a preliminary deblurring image, constraining is carried out through a plurality of regularization items, the prior distribution characteristics of a noise image gradient can be utilized to carry out gradient-related distributed regulation processing on a regularization weight factor, good image edge characteristics can be kept while noise can be effectivelyinhibited, and an image with high quality is obtained.