摘要
结合ROF模型和四阶PDE去噪模型提出一种改进的分数阶自适应PDE去噪算法。改进算法根据各个像素点的梯度信息自适应地选择分数阶PDE的阶数,根据尺度参数的特点,提出一种自适应的尺度参数迭代算法。数值实验表明,改进算法能够较好地提高峰值信噪比,保护边缘,有效抑制"阶梯效应"。
An improved fractional adaptive PDE denoising algorithm based on the ROF model and the fourth-order PDE denoising model is presented. A selection function of the fractional PDE＇s order is proposed. The selection function adopts adaptive order at different pixels of the image. In addition, an adaptive iterative algorithm of the scale parameter is put forward. Numerical experiments show that the novel algorithm can increase the PSNR, preserve edges and restrain the ＂step effect＂ effectively.
出处
《河北科技大学学报》
CAS
北大核心
2011年第6期580-584,604,共6页
Journal of Hebei University of Science and Technology
基金
河南省重点科技攻关项目(082100210041)
河南省自然科学基础研究计划项目(2007140020)
关键词
图像去噪
全变分
分数阶导数
偏微分方程
image denoising
total variations fractional derivative
partial differential equation
