Ref:
A Python article I received from one of the mailing lists I subscribe to.
Viz.:
Assuming you have the bandwidth to read it, this is an interesting article about Fourier transfoms, how they work, and what kind of stuff they’re useful for.
My first-pass takeaway from this is if you have something, (video, audio, radio, image data, etc.), that can be classified into distinct parts, a Fourier transform can help isolate specific parts of the signal/image.
For example:
Say you want to do something like lane following, and the lane is clearly marked on each side with some kind of higher contrast paint.
You could examine the image with a Fourier transform, isolating the two frequency strokes representing the edges of the road, and if they are equal, you’re centered.
Obviously a tremendous over simplification, but you should get the point.
It’s well worth the read.
What say ye?