1.Mat矩阵数值的存储方式
这里以指针的方式访问图像素为例
(1)单通道
定义一个单通道图像：
cv::Mat img_1 = (320, 640, CV_8UC1, Scalar(0));对于单通道M(i,j)即为第i行j列的其灰度值；程序中表示为：
img_1.ptr<uchar>(i)[j];(2)多通道
这里以RGB图像为例，每一个子列依次为B、G、R，，第一个分量是蓝色，第二个是绿色，第三个是红色。
定义一个3通道BGR图像：
cv::Mat img_1 = (320, 640, CV_8UC3, Scalar(0， 0 ，0));对于多通道M(i,j*3)即为第i行j列的B通道其灰度值，M(i,j*3+1) 即为第i行j列的G通道其灰度值，M(i,j*3+1) 即为第i行j列的B通 道其灰度值；程序中表示为：
第i行j列的B通道其灰度值：
img_1.ptr<uchar>(i)[j*3];第i行j列的G通道其灰度值：
img_1.ptr<uchar>(i)[j*3+1]; 第i行j列的R通道其灰度值：
img_1.ptr<uchar>(i)[j*3+2];2.示例程序，以三种方法（指针，at，迭代器）
获得图像像素值#include <iostream>#include <vector>#include <algorithm>#include <opencv2/core/core.hpp>#include <opencv2/imgproc/imgproc.hpp>#include <opencv2/opencv.hpp>using namespace std;using namespace cv;void get_setImagePixel3(char *imagePath, int x, int y){Mat image = imread(imagePath, 1);//得宽高int w = image.cols;int h = image.rows;int channels = image.channels();if (channels == 1){//得到初始位置的迭代器 Mat_<uchar>::iterator it = image.begin<uchar>();//得到终止位置的迭代器 Mat_<uchar>::iterator itend = image.end<uchar>();int pixel = *(it + y * w + x);cout << "灰度图像,处的灰度值为" << pixel << endl;}else{//得到初始位置的迭代器 Mat_<Vec3b>::iterator it = image.begin<Vec3b>();//得到终止位置的迭代器 Mat_<Vec3b>::iterator itend = image.end<Vec3b>();//读取it = it + y * w + x;int b = (*it)[0];cout << b << endl;int g = (*it)[1];cout << g << endl;int r = (*it)[2];cout << r << endl;//设置像素值(*it)[0] = 255;(*it)[1] = 255;(*it)[2] = 255;}imshow("cc", image);}int main(){vector<int> v = {1,2,3,4,6};cout << "*********通过指针访问像素的灰度值********************" << endl;//通过指针访问像素的灰度值//单通道Mat img1(20, 30, CV_32FC1, Scalar(0));img1.ptr<float>(19)[25] = 23456.1789;cout << "img(19,25):" << img1.ptr<float>(19)[25] <<endl;//多通道,创建一个B、G、R通道灰度值为0的图像，图像大小20*30，每个灰度值为16位无符号2进制表示Mat img2(20, 30, CV_16UC3, Scalar(0, 0, 0));cout << "img(1,2):" << int(img1.ptr<uchar>(19)[25]) << endl;Mat img = imread("test1.jpg");int numRow = img.rows;int numCol = img.cols;int numCol_channel = img.cols*img.channels();cout << "numRow:" << numRow << endl;cout << "numCol:" << numCol << endl;cout << "numCol_channel:" << numCol_channel << endl;cout << "45行，483列B通道灰度值" << int(img.ptr<uchar>(45)[483*3]) << endl;cout << "45行，483列G通道灰度值" << int(img.ptr<uchar>(45)[483*3+1]) << endl;cout << "45行，483列R通道灰度值" << int(img.ptr<uchar>(45)[483*3+2]) << endl;Mat img_B(numRow, numCol, CV_8UC3, Scalar(0, 0, 0));Mat img_G(numRow, numCol, CV_8UC3, Scalar(0, 0, 0));Mat img_R(numRow, numCol, CV_8UC3, Scalar(0, 0, 0));for (int i = 0; i < numRow; i++){for (int j = 0; j < numCol; j++){img_B.ptr<uchar>(i)[j*3] = img.ptr<uchar>(i)[j*3];img_G.ptr<uchar>(i)[j*3+1] = img.ptr<uchar>(i)[j*3+1];img_R.ptr<uchar>(i)[j*3+2] = img.ptr<uchar>(i)[j*3+2];}}imshow("img", img);imshow("img_B", img_B);imshow("img_G", img_G);imshow("img_R", img_R);cout << endl;cout << endl;cout << endl;cout << "*********at只适合灰度值为8位的图像********************" << endl;//注意：at只适合灰度值为8位的图像//单通道Mat img3(20, 30, CV_8UC1, Scalar(0));cout << "img（7,8）" << int(img3.at<uchar>(7, 8)) << endl;//多通道Mat img4(20, 30, CV_8UC3, Scalar(0));//BGR通道cout << "B通道灰度值" << int(img4.at<Vec3b>(3, 4)[0]) << endl;cout << "G通道灰度值" << int(img4.at<Vec3b>(3, 4)[1]) << endl;cout << "R通道灰度值" << int(img4.at<Vec3b>(3, 4)[2]) << endl;waitKey(0);system("pause");return 0;}
函数)
