Course link:
5.fourier transform (Fourier transform)_bilibili_bilibili
1. Effect (time domain image——>frequency domain image)
Natural images: low-frequency information (brighter) > high-frequency information
——>
(very mushy and flat)
After Fourier transform, the four corners should be brighter and the middle should be darker. The bright spots in the four corners are moved to the middle, causing the middle to be brighter and the surroundings to be darker.
Only the bright area in the middle is left, and the remaining parts are set to 0 (only low-frequency information is retained), and then the inverse (inverse) Fourier transform is performed on it, and a very blurry and flat picture is obtained.
(It’s all about the details, and the places that have changed a lot have been preserved)
Only the dark ones (high-frequency information) are retained, and then the inverse Fourier transform is performed, and the result is a picture like this.
Frequency domain filtering: Time domain image ——> Frequency domain image (processed) ——> Time domain image (magical changes occur)
2. What is frequency domain?
Periodic function: can be divided into many superpositions of sine waves and cosine waves.
In fact, it is a function of amplitude with respect to frequency.
The origin of Fourier transform:
Fourier transform algorithm (DFT, FFT)
