摘要
为了抑制更好的抑制噪声保留边缘信息,提出了一种各向异性高斯滤波的改进方法,该方法先用中值滤波去除椒盐噪声,再利用灰度共生矩阵的惯性矩特征值调整滤波器的长轴尺度,实现各向异性高斯滤波.实验证明,本文所提出的改进方法,在噪声抑制和边缘保持方面优于各向异性高斯滤波器.
This paper proposes an improved bit-allocation algorithm for anisotropic Gaussian filter to smooth a noisy image without blurring the edges.A median filter is used to smooth salt and pepper noise,then the variance of gray-scale is used to adjust Gaussian space coefficient,so the image edge can be automatically detected.The results indicate that the image edge can be well preserved and the noise can be got rid of completely through this improved algorithm.
出处
《西南民族大学学报（自然科学版）》
CAS
2007年第1期120-123,共4页
Journal of Southwest Minzu University(Natural Science Edition)
基金
国家自然科学基金(60272060)
四川省青年科技基金(05ZQ026-019)资助项目
关键词
边缘检测
各向异性高斯滤波
中值滤波
惯性矩特征
edge detection
Anisotropic Gaussian filter
median filter
variance of gray-scale
