基于梯度正则化与自适应可信度的二阶PDE去噪模型

A Second-Order PDE Denoising Model Based on Gradient Regularization and Adaptive Credibility

针对抑制图像过程中出现边缘模糊的问题,提出一种基于梯度正则化与自适应可信度的二阶偏微分图像去噪模型。首先,针对现有的二阶偏微分方程(Partial Differential Equation,PDE)去噪模型的缺点,提出一个新的扩散函数。该模型能够根据图像局部特征的差异来调整图像扩散系数。其次,针对二阶PDE模型的保真项系数的特点,将保真项的系数由常数转变为函数,以达到平衡平滑区域和纹理区域的效果。最后,利用模型实现图像去噪的数值。仿真实验结果表明,改进的二阶PDE模型在去噪的同时较好地保留了图像的纹理细节,是一种有效的去噪算法。

Aiming at the problem of blurred edge information in the process of suppressing image noise, this paper proposes an image denoising method based on an improved second-order partial differential model. Firstly, a new diffusion function is proposed to overcome the shortcomings of existing second-order Partial Differential Equation(PDE)denoising models. The model can adjust the image diffusion coefficient according to the difference of image local features.Next, according to the characteristics of the fidelity term coefficients of the second-order PDE model, the fidelity term coefficients are transformed from constants to functions to achieve the effect of balancing smooth regions and texture regions. The simulation results show that the improved PDE model is a more effective algorithm.