新型梯度倒数加权滤波器及其性质

New Gradient Inverse Weighting Filter and Its Properties

该文从经典梯度倒数加权滤波器出发 ,通过重新定义合适的模板结构和参数提升梯度模 ,提出了一种应用于信号去噪和图像增强的新型梯度倒数加权滤波器。分析该滤波器的迭代过程 ,揭示了该数字滤波器解决了整体变分的全局最优解问题。该文既以实验说明了该滤波器优良的去噪和边缘保持能力 。

Based on the classical gradient inverse weightin g filter, a new gradient inverse weighting filter for signal denoising a nd image en hancement is proposed. In the filter, the filter stencil and the parameter lifti ng gradient modulus is defined. In iterations process, it solves a global total variational optimiza tion problem. The feasibility of the filter in de noising and edge prese rving is demonstrated with experiments, and some results in theory between the filter and total variation model are proved in mathematically.