OpenCV白平衡算法之灰度世界法（消除RGB受光照影响）

　　在用OpenCV对图像进行处理时，利用颜色定位是常常会接触到的方法，但RGB受光照影响比较严重，转换到HSV XYZ等空间也解决不了时，

可以用白平衡算法进行修正，使其发黄、发蓝、发红的照片更加趋于自然光下的图像。（转摘请说明来源）

程序代码示例如下：

 1 //该代码实现白平衡算法中的灰度世界法，能有效改善图像发红发蓝发绿的现象；
 2 
 3 #include 
 4 using namespace cv;
 5 
 6 int main()
 7 {
 8     Mat g_srcImage,dstImage;
 9     vector g_vChannels;
10     g_srcImage = imread("C:/Users/Administrator/Desktop/区分高架定位/01.jpg");
11     imshow("原图",g_srcImage);
12     //waitKey(0);
13 
14     //分离通道
15     split(g_srcImage,g_vChannels);
16     Mat imageBlueChannel = g_vChannels.at(0);
17     Mat imageGreenChannel = g_vChannels.at(1);
18     Mat imageRedChannel = g_vChannels.at(2);
19 
20     double imageBlueChannelAvg=0;
21     double imageGreenChannelAvg=0;
22     double imageRedChannelAvg=0;
23 
24     //求各通道的平均值
25     imageBlueChannelAvg = mean(imageBlueChannel)[0];
26     imageGreenChannelAvg = mean(imageGreenChannel)[0];
27     imageRedChannelAvg = mean(imageRedChannel)[0];
28 
29     //求出个通道所占增益
30     double K = (imageRedChannelAvg+imageGreenChannelAvg+imageRedChannelAvg)/3;
31     double Kb = K/imageBlueChannelAvg;
32     double Kg = K/imageGreenChannelAvg;
33     double Kr = K/imageRedChannelAvg;
34 
35     //更新白平衡后的各通道BGR值
36     addWeighted(imageBlueChannel,Kb,0,0,0,imageBlueChannel);
37     addWeighted(imageGreenChannel,Kg,0,0,0,imageGreenChannel);
38     addWeighted(imageRedChannel,Kr,0,0,0,imageRedChannel);
39 
40     merge(g_vChannels,dstImage);//图像各通道合并
41     imshow("白平衡后图",dstImage);
42     waitKey(0);
43     return 0;
44 }

结果如下：

   

 

转载于:https://www.cnblogs.com/ggYYa/p/5707259.html