2D FFT of an image shows some invalid values for high frequencies

Question

Mona Mahboob Kanafi 2013-6-27

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/80412-2d-fft-of-an-image-shows-some-invalid-values-for-high-frequencies

I uploaded my 2D FFT magnitude image here:

https://dl.dropboxusercontent.com/u/82779497/2D_FFT.png

If you take a look at it, for high frequencies[right, left, top and bottom], only at around x and y axis, there are some points with high power[yellow color]. These points shouldn't be in the resultant FFT2, since I know the original height image is isotropic and therefore the 2D FFT must look something like the example below(just note high frequencies; they are not for the same image):

https://dl.dropboxusercontent.com/u/82779497/2D_PSD_fixed.png

Now, the question is, what could be the possible reasons for such a behavior at high frequencies?

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Matt J 2013-6-27

编辑：Matt J 2013-6-27

since I know the original height image is isotropic and therefore the 2D FFT must look something like the example below

Post the original height image, so that we can verify that it is isotropic.

Have you used 'axis equal' to display these images? I can discern some asymmetry in your 2nd FFT image, even though it is mostly isotropic.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Matt J 2013-6-27

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/80412-2d-fft-of-an-image-shows-some-invalid-values-for-high-frequencies#answer_90115

Now, the question is, what could be the possible reasons for such a behavior at high frequencies?

Because the frequencies you've shown are not really "high". Both plots look the same, except that the first one has smaller axis limits and is presenting a cropped view of the 2nd one.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Mona Mahboob Kanafi 2013-6-27

Hi Matt, by high frequencies I mean 'high' in comparison to zero. And those images are not the same, because the first one is an image of a real surface measurement but the second one is a theoretical generated surface to see how my code works. I asked the readers just to take a look at the second image on its high frequencies to see what I mean by isotropy in the image and what I don't want to see in the first image. And know whether it's a calculation problem or some problem in the profilometer or the software we use while exporting the data.

请先登录，再进行评论。

Answer 2

Matt J 2013-6-27

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/80412-2d-fft-of-an-image-shows-some-invalid-values-for-high-frequencies#answer_90120

编辑：Matt J 2013-6-27

Is it possible that you have non-isotropic noise? Even if the underlying input signal is isotropic, asymmetric high freqeuncy noise could certainly cause asymmetry in the high frequency spectrum.