every blog every motto: You can do more than you think.
https://blog.csdn.net/weixin_39190382?type=blog

0. 前言

访问图像元素

1. 访问图像像素

1.1 访问某像素

//灰度图像：
image.at<uchar>(j, i) //j为行数，i为列数
//BGR彩色图像
image.at<Vec3b>(j, i)[0] //B分量
image.at<Vec3b>(j, i)[1] //G分量
image.at<Vec3b>(j, i)[2] //R分量

1.2 遍历像素

以添加噪声为例

1.2.1 准备

创建添加噪声函数

Salt.h

#pragma once
#include <iostream>
#include <opencv2/opencv.hpp>
#include <random>

using namespace cv;
using namespace std;

void Salt(Mat img,int n); // n加入噪声点数

Salt.cpp

#include "Salt.h"

void Salt(Mat img,int n){
    // 随机数生成器
    default_random_engine generater;
    uniform_int_distribution<int> randRow(0,img.rows-1);
    uniform_int_distribution<int> randCol(0,img.cols-1);

    int i,j;
    for (int k=0;k<n;k++){
        i=randRow(generater);
        j=randCol(generater);

        if (img.channels() == 1){
            img.at<uchar>(i,j) = 255;
        }
        else if (img.channels() == 3){
            img.at<Vec3b>(i,j)[0] = 255;
            img.at<Vec3b>(i,j)[1] = 255;
            img.at<Vec3b>(i,j)[2] = 255;
        }
    }
}

1.2.2 添加噪声

#include "Salt.h"
#include <iostream>
#include <opencv2/opencv.hpp>


using namespace cv;
using namespace std;


int main(){
    Mat img = imread("/home/v/home.png");
    if (img.empty()){
        cout<<"Could not open or find the image"<<endl;
        return -1;
    }

    imshow("Img",img);

    Salt(img,5000); // 加入噪声点
    imshow("Salt",img);
    waitKey(0);
    return 0;
}

1.3 指针遍历

1.3.1 以卷积运算为例

#include "Salt.h"
#include <iostream>
#include <opencv2/opencv.hpp>

using namespace cv;
using namespace std;


int main(){
    Mat img = imread("/home/v/home.png");
    if (img.empty()){
        cout<<"Could not open or find the image"<<endl;
        return -1;
    }

    Mat out_img;
    out_img = Mat(img.size(),img.type()); // 定义输出图像大小
    out_img = img.clone(); // clone原图像像素值

    int rows = img.rows; // 原图行数
    int stepx = img.channels(); // 原图通道数
    int cols = (img.cols)*img.channels(); // 矩阵总列数

    for (int row=1;row<rows-1;row++){
        const uchar* previous = img.ptr<uchar>(row-1); // 原图上一行指针
        const uchar* current = img.ptr<uchar>(row); // 原图当前行指针
        const uchar* next = img.ptr<uchar>(row+1); // 原图下一行指针

        uchar* output = out_img.ptr<uchar>(row); // 输出图像当前指针

        for (int col=stepx;col<cols-stepx;col++){ // 对列进行遍历
            // saturate_cast<uchar>(a) 当a小于0，输出0，当a大于255输出255，0-255之间原样输出
            output[col] = saturate_cast<uchar>(5*current[col]-(previous[col]+current[col-stepx] + current[col+stepx] + next[col]));
        }
    }
    imshow("Img",img);
    imshow("Out",out_img);

    waitKey(0);
    return 0;
}

1.3.2 自带的卷积运算

#include "Salt.h"
#include <iostream>
#include <opencv2/opencv.hpp>


using namespace cv;
using namespace std;


int main(){
    Mat img = imread("/home/v/home.png");
    if (img.empty()){
        cout<<"Could not open or find the image"<<endl;
        return -1;
    }
    Mat out_img;

    Mat kernel = (Mat_<char>(3,3)<<0,-1,0,-1,5,-1,0,-1,0);
    filter2D(img,out_img,img.depth(),kernel); // 卷积

    imshow("Img",img);
    imshow("Out",out_img);
    waitKey(0);
    return 0;
}