Suppose a sequence of numbers (wave-like data). Then I do the DFT (or FFT) conversion. The next step I want to achieve is to search for frequencies corresponding to the real frequencies that are included in the data. As you know, the DFT output has the real and imaginary part a [i] and b [i]. If you look at the spectrum (sqrt (a [i] ^ 2 + b [i] ^ 2), then the maximum in it corresponds to the frequency that is included in the data. The question is, how to find all frequencies from the DFT? The problem arises when there is many other peaks that can be falsely selected.
I had a similar problem when processing spectral data analysis when I was writing my thesis.
: .
, - , (). , DFT . , . DFT , , , / .
, , (), (). () . , , , (), , .
, , . , , .
, , .
DFT, . , n , dt, , 5 (), 1, 1/(n * dt) . ( , Pi - )