"Digital Image Processing" third edition Rafael C. Gonzalez waiting for P162
This chapter is also relatively simple. I understand it, that is, convert the picture f(x,y) into a Fourier transform F(u,v), and then modify F(u,v). This can be changed by a function H(u,v) and F(u,v) array complex multiplication,
Of course, you can also set it directly, such as F(1,2)=1+4i.
After setting, turn F(u,v) back with IDFT to g(x,y) and display it as a picture effect
This is frequency domain filtering
First, the host experiment
(1) Ideal low-pass filter ILPF, the fuzzy picture effect of different D0 situations, 160 is already very close to the original picture
https://github.com/penkee/imagecal/blob/master/app-dao/src/test/java/com/dcloud/app_dao/TestDFT.java
frequencyFilter_ ILPF () method tested
(2) Gaussian low-pass filter
In the follow-up update, in addition, the host has worked hard for more than a month to update all DFTs into fast transforms, that is, FFT. The speed is super fast, 256 pictures are completed in 90ms.