How to determine the radius of gaussian smoothing?

9 次查看（过去 30 天）

hanif hamden 2020-12-4

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/678313-how-to-determine-the-radius-of-gaussian-smoothing

评论： hanif hamden 2020-12-4

Example.txt

Hi everyone, can anyone explain how I would like to determine the radius smoothing by using imgaussfilt function matlab.

Let say, in my coding, I try imgaussfilt function with 6 sigma. Since the default filter size is 2*ceil(2*sigma)+1.

Thus, the filter size for 6sigma is 2*2*6+1 = 25 . So, would the smoothing radius is half of the filter size? correct me I'm wrong. Below is my coding. Thank you.

clc;clear all; close all;
file = 'Example.txt';
name = 'Testing Smoothing';
Data = textread(file);
y = Data(:,1);
x = Data(:,2);
z = Data(:,3);
ymin = 0.5 ; ymax = 3.5;
xmin = 96 ; xmax = 123;
xlim = [xmin xmax];
ylim = [ymin ymax];
m2 = 0.033333333333; 
    
dx =xmin:m2:xmax;   % set the min,max y and interval
dy = ymin:m2:ymax; % set the min,max x and interval
[Xi,Yi] = meshgrid(dx,dy); 
[Zi] = griddata(x,y,z,Xi,Yi,'cubic');
Zi = imgaussfilt(Zi,6,'FilterDomain','spatial');
GrdXYZ = [Yi(:) Xi(:) Zi(:)];
nGrd = sortrows(GrdXYZ,1);

2 个评论
显示无隐藏无

Steve Eddins 2020-12-4

What is your definition of smoothing radius for a Gaussian filter? Theoretically speaking, the Gaussian function has infinite extent. Practically, it decays to zero quickly so that a finite-extent window is sufficient. But it's not clear to me whether smoothing radius is clearly defined for it.

hanif hamden 2020-12-4