C++ - Convert Mat to Opencv integer array?

Question

I’m doing an image manipulation project using the Opencv. I need to analyze the values of pixels of a "binarized" image, for this I am trying to convert my Mat file image to a array of integers, more specifically an array of integers. I am trying to do as follows:

Mat m;
m = imread("C:/Users/Syn/Desktop/filtesteBranco.jpg");

int matriz[this->m.size().width][this->m.size().height]; 
int i, j; 

// inicializar matriz com zeros //
for (i = 0; i < (m.rows); i++){ 
    for (j = 0; j <(m.cols); j++) 
    {
        matriz[i][j] = 0;
    }
}

for (int x = 0; x < (m.rows); x++) // varredura da matriz
{
    for (int y = 0; y<(m.cols); y++)
    {
        matriz[x][y] = m.at<uchar>(x,y);    // capturando os pixels
    }
}     

However, when trying to display the array and compare its integer values to the image, the values do not match the appropriate pixels on the lines of the original image. I’m capturing the values of pixels the wrong way? Would anyone know the right way to do it? Thanks in advance.

EDIT:

Below is the test I’m doing, I added 3 white pixels at the beginning and end of the last line of the image. The image in question has size 674x35, so I posted two zoom images to show what I’m doing.

When printing the values of the last line of the image, I got the following values, however, they do not match the pixels of the image:

The initial and final values of the line are completely wrong. Someone would know what might be happening?

Answer 1

The iteration over the pixels of an image with Mat is made the way you already use it. Here is an example program:

#include "opencv2/core.hpp"
#include "opencv2/opencv.hpp"

using namespace cv;

int main(int argc, char**argv)
{
    Mat m;
    m = imread("C:/temp/teste.bmp", 0);

    for(int x = m.rows-1; x < (m.rows); x++)
    {
        for(int y = 0; y < (m.cols); y++)
        {
            printf("%d ", m.at<uchar>(x, y));
        }
    }

    return 0;
}

This program, reading the following image:

(enlarged:)

Results in the following:

0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255

But note that the image, being "binary", needs to be recorded as BMP (Bit Map), so that the pixel values are directly recorded and not any compressed format. Also note that the call from imread should read the image in a single channel (so the second parameter as 0).

If you use JPG, it can generate "pixels" with details missing due to the compression used: