摘要
提出一种新的去除高斯噪音的方法(NLTF)。该算法利用梯度能很好地反映图像结构信息的特点。把梯度引入双边滤波算法中,又结合非局部均值的思想,而提出一种改进算法。它在保护图像边界和细节方面比双边滤波算法有明显优势。在去除高斯噪音上比非局部均值平滑得更好。仿真实验证明,使用NLTF去噪,在视觉效果和PSNR等方面均超过已知的许多经典算法。
A new method for removal of Gaussian noise （NLTF） is proposed. It uses the gradient can well re- flect the characteristics of the image structural information. Gradient is introduced in the bilateral filtering algo- rithm, also combines the idea of non-local mean, and proposed an improved algorithm. It has the obvious advantage in protecting the boundaries and detail than the bilateral filtering algorithm. It is smooth noise better than the non-local mean in the removal of Gaussian noise. Simulation results show, NLTF denoising, it is over known classical algorithm in visual effects and PSNR.
出处
《科学技术与工程》
北大核心
2012年第24期6204-6207,共4页
Science Technology and Engineering
基金
河南省教育厅自然科学基金项目(2009C110004)
河南省高校青年骨干教师项目(2009GGJS-176)资助
关键词
高斯噪音
滤波算法
梯度
Gaussian noise filtering algorithm gradient
