DFT Frequency Components Opencv

问题:

I used the following link to learn about how to use DFT in Opencv

    http://docs.opencv.org/doc/tutorials/core/discrete_fourier_transform/discrete_fourier_transform.html

I understood how the magnitude is extracted from the Dft. However, I want to know for what frequencies each magnitude stands for, to know about the presence of high frequencies and low frequencies. Could you please help me how to intepret this? For what frequencies each magnitude is coefficient for?

I want to know this without plotting, as I want to use this data autonomously, without manually referring to from the plot. Please help me

回答1:

Sounds like you need a signal processing lesson instead of a computer vision lesson. What you get from the DFT is a matrix of complex components as big as the image you put into it. These correspond to the frequency components from 0 (top left) to the sampling frequency (bottom right). A component with frequency equal to the sampling frequency is a component with a period of 1 pixel. A component with a horizontal and vertical period of 4 pixels has a frequency of a quarter of the sampling frequency, so can be found at position [rows/4, cols/4], since four times longer period means four times smaller frequency.

Say you are looking for the component with horizontal period of 10 pixels and vertical period of 6 pixels. This component can be found at position [rows/6, cols/10] in the DFT result.