Opencv中RGB通道/HSV通道并分离

向右看齐 2022-01-09 04:41 360阅读 0赞

> [ OpenCV中HSV颜色模型及颜色分量范围][OpenCV_HSV]
> 
> [opencv HSV 颜色模型(H通道取值 && CV\_BGR2HSV＿FULL)][opencv HSV _H_ _ CV_BGR2HSV_FULL]
> 
> [【opencv】在hsv颜色空间识别区域颜色][opencv_hsv]

将原图分离为RGB单通道和B三通道（GR通道均为0）

1     Mat src_color = imread(path);
     2     vector<Mat> channels;
     3     vector<Mat> mbgr(3);
     4     split(src_color, channels);
     5     Mat B = channels.at(0);
     6     Mat G = channels.at(1);
     7     Mat R = channels.at(2);
     8     Mat bk1(src_color.size(), CV_8UC1, Scalar(0));
     9     Mat imageB(src_color.size(), CV_8UC3);
    10     mbgr[0] = B;
    11     mbgr[1] = bk1;
    12     mbgr[2] = bk1;
    13     merge(mbgr, imageB);
    14     imshow("imageB", imageB);
    15     imshow("原图", src_color);
    16     imshow("B", B);
    17     imshow("G", G);
    18     imshow("R", R);
    19     waitKey(0);

![1121288-20180131225142468-1945035426.png][]

![1121288-20180131225123062-1149054164.png][]

![1121288-20180131225226375-528138981.png][]

![1121288-20180131225254078-829746309.png][]

![1121288-20180131225500984-467303709.png][]

利用opencv把RGB图片像HSV颜色空间转变的时候，H通道的值范围为： 0-180，S： 0-255，V：0-255。

利用直方图显示三个通道的时候，H通道都集中在前半部分，想让它取值范围扩大。后来发现了 CV\_BGR2HSV＿FULL ，CV\_BGR2HSV 在转换图像的时候是将 H / 2 ---> H ，我们知道图像中色相H的取值范围为 0-360 ，所以利用opencv转换之后得到的H的范围为 0-180。而 CV\_BGR2HSV＿FULL 实现的映射是 H \* 255 / 360 --->H , 所以利用\_FULL 这个转换得到的H通道图像的范围为 0-255这两种方式都是opencv把\[0-360\] 的数据压缩到一个字节可以处理的数据。**\[ 所以这里说的图像的depth = 8， 刚好能表示的最大值是 255\]**

1 Mat src_hsv;
     2 cvtColor(src_color, src_hsv, CV_BGR2HSV_FULL);
     3 //cvtColor(src_color, src_hsv, CV_BGR2HSV);
     4 vector<Mat> hsv_channels;
     5 split(src_hsv, hsv_channels);
     6 Mat H = hsv_channels.at(0);
     7 Mat S = hsv_channels.at(1);
     8 Mat V = hsv_channels.at(2);
     9 imshow("H", H);
    10 imshow("S", S);
    11 imshow("V", V);

![1121288-20180202143556640-1292729975.png][]

![1121288-20180202143615765-1763695779.png][]

![1121288-20180202143633656-191184108.png][]

转载于:https://www.cnblogs.com/qinguoyi/p/8395271.html

[OpenCV_HSV]: https://www.cnblogs.com/wangyblzu/p/5710715.html
[opencv HSV _H_ _ CV_BGR2HSV_FULL]: http://blog.csdn.net/timidsmile/article/details/17297811
[opencv_hsv]: http://blog.csdn.net/xishanbeilu/article/details/51722736
[1121288-20180131225142468-1945035426.png]: /images/20220109/3fd0ec1f98494695824c22255a99a88d.png
[1121288-20180131225123062-1149054164.png]: /images/20220109/10665266e46746c2b56b6b44bcbe2ed7.png
[1121288-20180131225226375-528138981.png]: /images/20220109/ed25d6b9030741429107997f37f843f7.png
[1121288-20180131225254078-829746309.png]: /images/20220109/d0d96fb2c99b468c9429f37e0a51ce5e.png
[1121288-20180131225500984-467303709.png]: /images/20220109/e9672b3107d44a2ca028719ce18bd725.png
[1121288-20180202143556640-1292729975.png]: /images/20220109/a254f96ab6a24fa1a7684590960c74be.png
[1121288-20180202143615765-1763695779.png]: /images/20220109/89274915038c4b5486337f58184c1653.png
[1121288-20180202143633656-191184108.png]: /images/20220109/6120cb7249bd4455a8ef077d64677a64.png