在影象處理中，尤其是處理多通道影象時，有時需要對各個通道進行分離，分別處理；有時還需要對分離處理後的各個通道進行合併，重新合併成一個多通道的影象。opencv2和opencv3中實現影象通道的合併與分離的函式分別是cv::split()和cv::merge()。

1、多通道影象的分離

cv::split()的具體呼叫方法如下：

void cv::split(  
    const cv::Mat& mtx, //輸入影象  
    vector<Mat>& mv // 輸出的多通道序列（n個單通道序列）  
);

2、影象多個通道的合併

cv::merge()

void merge(  
    const vector<cv::Mat>& mv, // 輸入的多通道序列(n個單通道序列)  
    cv::OutputArray dst // 輸出影象，包含mv  
); 

程式碼示例如下：

#include <opencv2/opencv.hpp>
#include <vector>

using namespace std;

int main()
{
	cv::Mat src = imread("timg.jpg", cv::IMREAD_COLOR);
	cv::imshow("src", src);

	// Split the image into different channels
	std::vector<cv::Mat> rgbChannels(3);
	split(src, rgbChannels);

	// Show individual channels
	cv::Mat blank_ch, fin_img;
	blank_ch = cv::Mat::zeros(cv::Size(src.cols, src.rows), CV_8UC1);

	// Showing Red Channel
	// G and B channels are kept as zero matrix for visual perception
	std::vector<cv::Mat> channels_r;
	channels_r.push_back(blank_ch);
	channels_r.push_back(blank_ch);
	channels_r.push_back(rgbChannels[2]);

	/// Merge the three channels
	cv::merge(channels_r, fin_img);
	cv::imshow("R", fin_img);


	// Showing Green Channel
	std::vector<cv::Mat> channels_g;
	channels_g.push_back(blank_ch);
	channels_g.push_back(rgbChannels[1]);
	channels_g.push_back(blank_ch);
	cv::merge(channels_g, fin_img);
	cv::imshow("G", fin_img);


	// Showing Blue Channel
	std::vector<cv::Mat> channels_b;
	channels_b.push_back(rgbChannels[0]);
	channels_b.push_back(blank_ch);
	channels_b.push_back(blank_ch);
	cv::merge(channels_b, fin_img);
	cv::imshow("B", fin_img);

	cv::waitKey(0);
	return 0;
}