I am not very sure what is the definition of the frequency of an Image? Is that the change rate of the color so that the edge will have the higher frequency?
Also, what is the meaning of the horizontal and vertical white lines in the right image?
I'm assuming this is in grayscale because we're measuring on a scale from black to white. Is it possible to do this for colors, in which case how would the spectrum of a colored image look?
kc1
@Sleepyhead08 I believe in order to take the FT of a color image, you need to process each color channel independently, resulting in an output image for each color channel (3 total).
keenan
@Sleepyhead08 Yeah, @kc1 is right: usually to do a Fourier transform of an image you do it componentwise. Your choice of color space (which we'll talk about later) may have an impact here, but it's not a huge qualitative shift.
I am not very sure what is the definition of the frequency of an Image? Is that the change rate of the color so that the edge will have the higher frequency?
Also, what is the meaning of the horizontal and vertical white lines in the right image?
@tingp Here's a link to filters on images.
https://www.tutorialspoint.com/dip/high_pass_vs_low_pass_filters.htm
The frequency of an image is the 2-D Discrete FT of the image signal. For more intuitions I find this link useful:
https://photo.stackexchange.com/questions/40401/what-does-frequency-mean-in-an-image
I'm assuming this is in grayscale because we're measuring on a scale from black to white. Is it possible to do this for colors, in which case how would the spectrum of a colored image look?
@Sleepyhead08 I believe in order to take the FT of a color image, you need to process each color channel independently, resulting in an output image for each color channel (3 total).
@Sleepyhead08 Yeah, @kc1 is right: usually to do a Fourier transform of an image you do it componentwise. Your choice of color space (which we'll talk about later) may have an impact here, but it's not a huge qualitative shift.