The problem of noise removal after FFT of polygon mask

I want to perform FFT on the polygon mask. The polygon mask data is `hexagon.mat`, and the image `star` is obtained after the following operation.

Y_starburst =  abs(fftshift(fft2(hexagon))).^1.2;
Y_starburst = Y_starburst/255;
figure;
imshow(Y_starburst);

How to operate to get only six bright stripes and delete the surrounding messy stripes? Any help would be greatly appreciated.

Answer 1

[Matt J]

If you increase the sampling fineness, I expect the artifacts will diminish.

[zzzj1208]

Thank you very much for your reply.

I increased the number of sampling points to 2048, that is, the size of the hexagon is 2048x2048. After running the above code, the artifacts were not eliminated, as shown in the figure below.

[Matt J]

load hexagon

Y_starburst =  abs(fftshift(fft2(X))).^1.2;
Y_starburst = Y_starburst/255;
figure;
imshow(Y_starburst); caxis([0,255]); axis([181.8347  339.7194  181.5928  339.4775])

[zzzj1208]

Thanks again for your answer, looks closer to what I was looking for!

Because I want to approximate the diffraction pattern of light with a hexagonal FFT, using your above code can get `star pattern`.

But the constant problem is that there is some 'noise' that I don't want to think about (circled in red, sorry, I'm not sure how to say it in technical terms). Perhaps there is some way you can remove this noise so that it looks smoother and more like real world beams? as the `input_000042.png` shows. Any help would be greatly appreciated.