2D FFT for a non square masked image

[John Pierson]

Hi,
Basically i am trying to apply a mask to an bitmap RGB image and then carry out a FFT on the image.
It works on the whole 300x300 image but the masked area is adding towards the generated image - i believe!

So can a FFT be applied to a masked area that is non-square??
If so how?

John Pierson Lv 8.20

Masked image, mask and VI attatched.

nb, images have been converted to jpeg as .bmp is not alowed as an attachment.

[Nann]

Can not find the 2D Nyquist Shift VI and 2D Real FFT VI.  I guess they are from one of the old example posted many years ago when LabVIEW didnot support 2D FFT.

Now the FFT VI in LabVIEW can compute 2D FFT.  The FFT VI is a polymorphic VI.  If you connect 2D data to the FFT VI, it will use the 2D instance automatically.  The FFT VI also has some input to define whether to have Nyqust shit.  The FFT can also be applied on non-square 2D data.

Is it what you want? 

[John Pierson]

Yes, But i've tried and failed, is there any chance you could turn that into a VI??
What would i need to do to modify what i already have?
I'm running Lv 8.20

[John Pierson]

The newer bits i think appear in LV 8.5??
i've attatched one of the missing bits, but i can't seem to get the 2d fft.

Many thanks,

[rpursley8]

Here is one that I have.  It may be the same on you used.  I posted it a while back.  A 2D FFT is essentially taking a 1D FFT of each row, transposing the array and taking another 1D FFT of each row and the transposing it back.


Message Edited by rpursley8 on 05-30-2008 09:08 AM

[John Pierson]

Thanks for time on that one,
i'm a little confused about what your VI does, you feed it an array of data andit carries out the FFT in 1d, wthinin a for loop,
how did you remove the effect of the unwanted data on the rest of the data?
Sorry i'm a little slow when it comes to FFT.

Thanks for your continued patience!

[rpursley8]

A multidimensional FFT is an separable function, so it is valid to apply it in each direction sequentially.