Author(s)

Zhi FANG

Corresponding Author

Zhi FANG

ABSTRACT

With the continuous development of digital image processing technology, people pay more and more attention to the visual effect of images. We need and hope to get clear image quality, which requires perfect image denoising technology. The fractional partial differential equation difference method for image denoising is an important direction in the field of image denoising, and its numerical method research has important theoretical significance and practical value. This paper introduces two denoising models based on PDE, namely ROF model and LLT model. According to the comparison of these two models, a comprehensive model combining ROF model and LLT model with weight function is proposed. After analyzing the parallelism of solving the difference equation after discretization of the fourth-order PDE, the noisy image is divided into stripe overlapping data, and the image is denoised in parallel, which greatly reduces the running time.

KEYWORDS

Micro-partial equation, Image denoising, Global variational method