1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
|
function [img_cleaned]=striperemoval(img, varargin)
%STRIPEREMOVAL removes interference fringes/stripes from an image.
% [img_cleaned]=STRIPEREMOVAL(img, varargin)
% img - image with stripes to be removed
% varargin - not used, it is here for compatibility
% By Eugeniy E. Mikhailov 2014/11/25
%
% Be aware of "slow" spatial stripes
% i.e. stripes with characteristic width comparable to the Gaussian beam.
% Their Fourier components will overlap with Gaussian.
% Fringes appear like strong isolated peaks in the Fourier transformed image.
% What is even nicer they appear symmetrically with respect to origin but not symmetrical to
% the fold with respect to X or Y axes. See figure below
%
% ^
% |
% * |
% |
% oOo
% -------------O0O-------------->
% oOo <- Gaussian beam fft
% |
% | * <- stripes fft peak
% |
%
%
% So it easy to clean them, especially if their spatial frequency is high.
%
img_fourier=fftshift(fft2(img)); % move to Fourier space
imgfa= abs(img_fourier); % amplitude of Fourier components
imgfa= imgfa/max(imgfa(:)); % normalized amplitudes
[Ny, Nx] = size (img); % image size
mask = img_fourier.*0;
for xi = 1:Nx
% strictly speaking one of dimensions can be only half sampled due to symmetry of the mask
for yi = 1:Ny
xr = Nx - xi + 1; % reflected x
yr = Ny - yi + 1; % reflected y
mt = imgfa( yi, xi ); % test point
ms = imgfa( yr, xr ); % symmetrical point with respect to origin
mr = imgfa( yi, xr ); % symmetrical point with respect to Y axis (right)
md = imgfa( yr, xi ); % symmetrical point with respect to Y axis (down)
m = ( mt + ms ) - ( mr + md ); % high value to the points shown at figure
m = m/max([mt, ms, mr, md]);
mask(yi,xi) = m;
end
end
threshold = max(mask(:))/2;
mask = mask < threshold; % points due to stripes need to be removed
img_cleaned = abs( ifft2(ifftshift(img_fourier.*mask)) );
|
