Meaning of image FFT?

5 visualizaciones (últimos 30 días)
Matt Maccarone
Matt Maccarone el 20 de Mzo. de 2013
I am trying to decode a hidden image. When taking the FFT of the image, the plot shows peaks at normalized frequencies 0.5 and 0.8. Is this the information that I want to save or get rid of? The first image only had one frequency in it and I was able to use a low pass filter to unveil the hidden image.
I may have just answered my own question, but input would still be appreciated. The initial coded image is composed of many horizontal lines and many vertical lines. I need to remove each sets of these separately to show the hidden image. 0.5 and 0.8 would correspond to the frequencies that were used to create the horizontal and vertical lines?

Respuestas (1)

Image Analyst
Image Analyst el 21 de Mzo. de 2013
Spikes in the image spectrum indicate periodic structure in the spatial domain (your original image).
By filtering to keep the spikes, and then inverse transforming, you can recover the periodic structure.
By filtering to zero out the spikes, and then inverse transforming, you can recover the original information without the periodic structure.
It might work fairly well if your hidden image is periodic. This is basically "inverse filtering".

Categorías

Más información sobre Fourier Analysis and Filtering en Help Center y File Exchange.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by