首页|一种修正的脉冲噪声滤除算法

一种修正的脉冲噪声滤除算法

扫码查看
随着计算机技术的快速发展,图像去噪技术得到了迅速的发展.由于其高效性,图像去噪中的两阶段法引起了许多学者的关注.在两阶段法第二阶段的脉冲噪声去除中,针对原来的共轭梯度法充分下降性受参数选择限制的不足,提出了一种修正的共轭梯度(CG)算法.该方法的一个优点是每次迭代的搜索方向都满足充分下降条件,并且该条件不受参数选择限制.在Armijo线性搜索条件下,证明了其全局收敛性.数值实验结果表明,该方法滤除脉冲噪声不仅不受参数的影响,还能处理参数为负的情况,并且比原来的方法效果更好.
A Modified Algorithm for Impulse Noise Removal
With the rapid development of computer technology,the technology of image denoising has developed rapidly.In recent years,due to its high efficiency,the two-stage method in image denoising has gained considerable attention from many scholars.In the second phase of the two-phase method for impulse noise removal,aiming at the shortcoming of the restriction on the parameter section for sufficient descent condition of the original conjugate gradient method,a modified conjugate gradient(CG)method is proposed.A nice property of the scheme is that the search direction at each iteration satisfies the suffi-cient descent condition and this condition isn't limited by the parameter.Under the Armijo-type line search,its global convergence is proved.Some numerical results are shown that this methodnot only filters out pulse noise without being affected by parameters,but also handles cases where the parameter is nega-tive,and obtainspractical,robust performance in comparing with the original method.

image denoisingtwo-phase methodconjugate gradient methodglobal convergence

高佩婷、陈泽华、李毅飞、李馥彤、孟云霞

展开 >

太原理工大学计算机科学与技术学院(大数据学院),太原 030024

北方自动控制技术研究所,太原 030006

图像去噪 两阶段法 共轭梯度法 全局收敛

山西省青年基金国家青年基金资助项目

20220302121225562302331

2024

火力与指挥控制
火力与指挥控制研究会,火力与指挥控制专业情报网

火力与指挥控制

CSTPCD北大核心
影响因子:0.312
ISSN:1002-0640
年,卷(期):2024.49(9)