如何在opencv python中更改图像照明

Question

我在python opencv中读取图像，现在我需要将此图像上的照明更改为更暗或更亮，我应该使用哪种方法来启用此功能？

Answer 1

I know I am late, but I would suggest using gamma correction .

Now what is gamma correction ?

I will make it clear in layman's terms:

• To display image on a screen, input voltage is needed.
• This voltage is output as light intensity.
• In perfect world, input voltage would be linear to output intensity.
• But the real screen output is close to an exponential curve, the exponent being gamma .

Since the computer screen applies a gamma value to the image on screen, the process of applying inverse gamma to counter this effect is called gamma correction .

Here is the code for the same using OpenCV 3.0.0 and python:

import cv2
import numpy as np

def adjust_gamma(image, gamma=1.0):

   invGamma = 1.0 / gamma
   table = np.array([((i / 255.0) ** invGamma) * 255
      for i in np.arange(0, 256)]).astype("uint8")

   return cv2.LUT(image, table)

x = 'C:/Users/524316/Desktop/stack/test.jpg'  #location of the image
original = cv2.imread(x, 1)
cv2.imshow('original',original)

gamma = 0.5                                   # change the value here to get different result
adjusted = adjust_gamma(original, gamma=gamma)
cv2.putText(adjusted, "g={}".format(gamma), (10, 30),cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 0, 255), 3)
cv2.imshow("gammam image 1", adjusted)

cv2.waitKey(0)
cv2.destroyAllWindows()

Here is the original image:

Applying gamma of value 0.5 will yield:

Applying gamma of value 1.5 will yield:

Applying gamma of value 2.5 will yield:

Applying gamma of value 1.0 will yield the same image.

Code was borrowed from this link

Answer 2

I think you can done this with opencv. Here is my suggestion

import cv2
import numpy as np

img1 = cv2.imread('abc.jpg')
a = np.double(img1)
b = a + 15
img2 = np.uint8(b)
cv2.imshow("frame",img1)
cv2.imshow("frame2",img2)
cv2.waitKey(0)
cv2.destroyAllWindows()

Here i increased the brightness of image. If you use subtraction that will makes darker.

Answer 3

A small remark to complement Jeru Luke's answer. Be sure that both arrays are of type np.uint8 . The cv.LUT function name stands for "look-up-table". It means that each pixel from the image is replaced with a value from the table .

You could convert both arrays:

def adjust_gamma(image, gamma=1.0):
   invGamma = 1.0 / gamma
   table = np.array([
      ((i / 255.0) ** invGamma) * 255
      for i in np.arange(0, 256)])
   return cv2.LUT(image.astype(np.uint8), table.astype(np.uint8))

Or make sure that an image array is casted to the valid type before passing into adjust_gamma() function. It is easy to convert the image into float while applying various transformations and forget to restore valid type before adjusting gamma.