diff options
Diffstat (limited to 'striperemoval.m')
| -rw-r--r-- | striperemoval.m | 24 |
1 files changed, 14 insertions, 10 deletions
diff --git a/striperemoval.m b/striperemoval.m index 7563bf0..8c2f623 100644 --- a/striperemoval.m +++ b/striperemoval.m @@ -1,13 +1,17 @@ function [img_cleaned]=striperemoval(img, varargin) -% Cleans an image from the stripes produced by interference fringes. -% img - image with stripes to be removed -% varargin - not used, it is here for compatibility +%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 their Fourier components will overlap with Gaussian. -% Such parallel fringes appears like strong isolated peaks in the Fourier transformed image. -% What is even nice they appear symmetrical with respect to origin but not symmetrical to -% fold +% 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 % % ^ % | @@ -15,13 +19,13 @@ function [img_cleaned]=striperemoval(img, varargin) % | % oOo % -------------O0O--------------> -% oOo <- gaussian beam fft +% oOo <- Gaussian beam fft % | -% | * <- stripe fft peak +% | * <- stripes fft peak % | % % -% so it easy to clean them, especially if their spatial frequency is high. +% So it easy to clean them, especially if their spatial frequency is high. % img_fourier=fftshift(fft2(img)); % move to Fourier space |