IMAQ FFT Butterworth

[Thomas K.]

Hello

I am trying to do a Butterworth-Highpass on a FFT of an image with pixel-by-pixel calculation. It works, but it is very very low-speed. Is there a way to do this faster?

I generate an image which is multiplicated on the original FFT, the slowest part of it is the calculation of the Butterworth arbitrary function...

best regards

[SEgeler]

Hi Thomas K,

Can you upload the image and code snippet you use? How fast / slow is it?

Have a nice day,

[Thomas K.]

The image comes from a camera, about 1200x800 resolution. Processing is done after 1.5 minutes...

[SEgeler]

Hi Thomas K,

thanks for the screenshot. I actually hoped for a VI Snippet to insert it into LabVIEW if needed 🙂

I cannot directly see the cause why this code takes 1.5 minutes; but let me suggest some steps that will help reading and debugging the code later on

(1) Clean up the Block diagram. Data should always flow from left to right.

(2) Use IMAQ Dispose to deallocate memory you allocate using IMAQ Create.

(3) Do never ever use coercion dots going into a for loop. Why? LabVIEW does coerce statistically. You won't know how often your for loop works if you use this form of coercion! 🙂

I am pretty sure the loss of time is in one or perhaps two VIs. Please use the execution highlighting. I assume you will have 1 - 1.5 minutes wait time on one specific VI. Can you identify a "time thief"?