摘要
提出了一种多尺度形态金字塔扫描图像去噪算法,采用多尺度形态金字塔将图像进行分解,对第一级细节信号用形态梯度算子、HMT变换删除孤立噪声点;对其它细节信号形态开闭滤波,保留图像结构特征;最后利用处理后的细节信号重建图像。仿真结果表明,同中值滤波相比,算法能够更好地消除扫描图像的噪声,保留更多的细节信息。
A method of scanned image denoising based on mathematical morphological pyramid is proposed. It is used to decompose an image level -wise and the detail information is filtered in different resolution spaces, a cleaned image with the detail information processed is restructured. Experiment results demonstrate that the efficient of the proposal method, and the method, Compared with conventional median filter, is of the advantages in terms of eliminating scanned noise, fine edges preserving.
出处
《计算机工程》
CAS
CSCD
北大核心
2004年第20期136-137,共2页
Computer Engineering
关键词
数学形态学
形态金字塔
去噪
扫描图像
Mathematic morphology
Morphological pyramid
Denoising
Scanned image