Opencv image processing Laplace pyramid

Laplacian pyramid is mainly used to reconstruct images , Laplace is just to enlarge the image , You can predict the residuals , What is residual , When a small image is enlarged , You need to insert some pixel values , What is inserted directly above is 0, Laplacian pyramid algorithm can predict according to the surrounding pixels , So as to achieve the maximum restoration of the image .

I learned that the principle is as follows ： Using every layer of the Gaussian pyramid , Subtracting the prediction image after the upper layer image is sampled and Gaussian convoluted , A series of difference images are obtained LP Decompose the image （ among LP It's the Laplacian pyramid image ）.

There is also a formula for Laplace （ This is the first time in this series to write formulas ）, among L For the Laplacian pyramid image ,G For Gaussian pyramid images

import cv2 as cv
 
src = cv.imread("./lena.jpg")
print(src.shape[:2])
cv.imshow("src image", src)
#  Down sampling 1 Time 
PyrDown1 = cv.pyrDown(src)
print(PyrDown1.shape[:2])
cv.imshow("PyrDown1", PyrDown1)

#  Down sampling 2 Time 
PyrDown2 = cv.pyrDown(PyrDown1)
print(PyrDown2.shape[:2])
cv.imshow("PyrDown2", PyrDown2)
 
#  Sampling up 1 Time 
pyrUp1 = cv.pyrUp(PyrDown2)
print(pyrUp1.shape[:2])
cv.imshow("pyrUp1", pyrUp1)
 
#  Calculating Laplacian pyramid image 
#  sampling 1 Time  -  Sampling up 1 The next graph 
laplace = PyrDown1 - pyrUp1
cv.imshow("laplace", laplace)
cv.waitKey()