《基于二维伽马函数的光照不均匀图像自适应校正算法》python复现.

整个过程主要分为四步，一是将RGB颜色空间转换维HSV颜色空间，二是提取有光照分量，采用retinex相关算法对光照进行多尺度高斯卷积，三是采用伽马函数对多尺度光照分量进行处理，最后将HSV空间颜色转换为RGB。流程图见下，代码直接附上，尺度参数可调，伽马参数也可调。

import cv2
import numpy as np
import os
def adjust_gamma(image, gamma=1.0):
    invGamma = 1.0 / gamma
    # table = []
    print(image)
    image=np.array(image).astype(np.uint8)

    #     table.append(((i / 255.0) ** invGamma) * 255)
    # table = np.array(table).astype("uint8")
    gamma_table = [np.power(x / 255.0, invGamma) * 255.0 for x in range(256)]
    # numpy数组默认数据类型为int32，需要将数据类型转换成opencv图像适合使用的无符号8位整型uint8，否则会报错
    gamma_table = np.round(np.array(gamma_table)).astype(np.uint8)

    print(image)
    return cv2.LUT(image, gamma_table)


def MSR(img, scales):
 weight = 1 / 3.0
 scales_size = len(scales)
 h, w = img.shape[:2]
 log_R = np.zeros((h, w), dtype=np.float32)

 for i in range(scales_size):
  img = replaceZeroes(img)
  # print('img',scales[i],img)
  L_blur = cv2.GaussianBlur(img, (scales[i], scales[i]), 0)
  L_blur = replaceZeroes(L_blur)
  dst_Img = cv2.log(img / 255.0)
  dst_Lblur = cv2.log(L_blur / 255.0)
  dst_Ixl = cv2.multiply(dst_Img, dst_Lblur)
  log_R += weight * cv2.subtract(dst_Img, dst_Ixl)

 dst_R = cv2.normalize(log_R, None, 0, 255, cv2.NORM_MINMAX)
 log_uint8 = cv2.convertScaleAbs(dst_R)
 return log_uint8

def replaceZeroes(data):
 min_nonzero = min(data[np.nonzero(data)])
 data[data == 0] = min_nonzero
 return data
dir='H:/python_project/test/underwater/'
file=os.listdir(dir)
for name in file:
 imgname=dir+name
 # imgname='image/486.jpg'
 im =cv2.imread(imgname)

 H,W=im.shape[0:2]
 cv2.imshow('original image',im)

 HSV_img = cv2.cvtColor(im,cv2.COLOR_BGR2HSV)

 cv2.imshow('HSV_img',HSV_img)

 h_gray, s_gray, v_gray = cv2.split(HSV_img)
 # cv2.imshow('image1',h_gray)
 # cv2.imshow('image2',s_gray)
 # cv2.imshow('image3',v_gray)

 Hsize=min(H,W)+1
 print(Hsize)

 size = 3

 # q=np.sqrt(2)

 SIGMA1=15
 SIGMA2=31
 SIGMA3=61
 # print('q',q,[int(15/q), int(80/q), int(250/q)])
 scales = [15, 101, 301]
 gaus=MSR(v_gray, scales)
 # 二维gamma卷积
 m=np.mean(gaus)
 # out=np.zeros((H,W))
 gamma=np.power(0.8,((gaus-m)/m))
 print(gamma)
 out=255*(np.power(v_gray/255,gamma))
 out_frame = np.zeros((out.shape[0],out.shape[1],3),dtype = np.uint8)
 out_frame[:,:,0] = h_gray
 out_frame[:,:,1] = s_gray
 out_frame[:,:,2] = out
 cv2.imshow('out_frame',out_frame)
 RGB_img = cv2.cvtColor(out_frame,cv2.COLOR_HSV2BGR)
 cv2.imshow('gaus_image1',RGB_img)
 cv2.waitKey(0)
```python
在这里插入代码片