python opencv 实现Reinhard颜色迁移算法

 

Reinhard颜色迁移算法的过程很简单，流程如下，细节部分见原文，题目为color transfer between images：

1. 将参考图片和目标图片转换到LAB空间下
2. 得到参考图片和目标图片的均值和标准差
3. 对目标图片的每一个像素值，减去目标图像均值然后乘上参考图片和目标图片标准差的比值，再加上参考图像均值
4. 将目标图片转换到RGB空间

将RGB图片转换到LAB空间很重要，因为LAB空间能降低三原色之间的相关性，如果不转换，结果会有很大的不同

# -*- coding: utf-8 -*-


import cv2
import numpy as np
image = cv2.imread('des.jpg')
image = cv2.cvtColor(image,cv2.COLOR_BGR2LAB)
original = cv2.imread('src.jpg')
original = cv2.cvtColor(original,cv2.COLOR_BGR2LAB)


def getavgstd(image):　　　　//得到均值和标准差
    avg = []
    std = []
    image_avg_l = np.mean(image[:,:,0])
    image_std_l = np.std(image[:,:,0])
    image_avg_a = np.mean(image[:,:,1])
    image_std_a = np.std(image[:,:,1])
    image_avg_b = np.mean(image[:,:,2])
    image_std_b = np.std(image[:,:,2])
    avg.append(image_avg_l)
    avg.append(image_avg_a)
    avg.append(image_avg_b)
    std.append(image_std_l)
    std.append(image_std_a)
    std.append(image_std_b)
    return (avg,std)

image_avg,image_std = getavgstd(image)
original_avg,original_std = getavgstd(original)

height,width,channel = image.shape
for i in range(0,height):
    for j in range(0,width):
        for k in range(0,channel):
            t = image[i,j,k]
            t = (t-image_avg[k])*(original_std[k]/image_std[k]) + original_avg[k]
            t = 0 if t<0 else t
            t = 255 if t>255 else t
            image[i,j,k] = t
image = cv2.cvtColor(image,cv2.COLOR_LAB2BGR)
cv2.imwrite('out.jpg',image)

结果如下：