Implementing For loop in open cv? - c++

Question

I want to implement similar loop in open cv.This code is done in Matlab .As i am new to open cv.I dont know how to proceed.Can Anyone give me ideas to do this in C++
for m=1:10
for n=1:20
for l=1:Ns
for k=1:Ns
Y(l,k)=image1(m-Ns+l-1,n-Ns+k-1);
DD(l,k)=image2(m-Ns+l-1,n-Ns+k-1);
end
end
e=Y-DD ;
end
end
here Image1 and image2 are 300*300 pixel in size. Y ,DD,image1,image2 al are mat images.

Answer 1

In OpenCV, the images can be represented as either Mat or IplImage. Your question does not specify the type of the image.
If IplImage:
IplImage *img;
unsigned char *image = (unsigned char*)(img->imageData);
int imageStride = img->widthStep;
pixData = image[xCount + yCount*imageStride];
If Mat:
Mat img;
unsigned char *image = (unsigned char*)(img.data);
int imageStride = img.step;
pixData = image[xCount + yCount*imageStride];
pixData will contain that data at (xCount, yCount). You could use this understanding in the for loop.
As you already know the logic, I am mentioning only on how to access data from a particular point in an image.

Answer 2

The most efficient way in OpenCV to access pixels in a for loop is:
cv::Mat rgbImage;
cv::Mat grayImage;
for ( int i = 0; i < rgbImage.rows; ++i )
{
const uint8_t* rowRgbI = rgbImage.ptr<uint8_t> ( i );
const uint8_t* rowGrayI = grayImage.ptr<uint8_t> ( i );
for ( int j = 0; j < rgbImage.cols; ++j )
{
uint8_t redChannel = *rowRgbI++;
uint8_t greenChannel = *rowRgbI++;
uint8_t blueChannel = *rowRgbI++;
uint8_t grayChannel = *rowGrayI++
}
}
Depending whether your image is one or more channels you can modify the code above.
If you want to implement window sliding, you could do something like this:
cv::Mat img;
int windowWidth = 5;
int windowHeight = 5;
for ( int i = 0; i < img.rows - windowHeight; ++i )
{
for ( int j = 0; j < img.cols - winddowWidth; ++j )
{
// either this
cv::Mat currentWindow = img(cv::Range(j, i), cv::Range(j + windowWidth, i + windowHeight));
// perform some operations on the currentWindow
// or do this
getRectSubPix(img, cv::Size(windowWidth, windowHeight), cv::Point2f(j, i), currentWindow));
// perform some operations on the currentWindow
}
}
You can read more about getRectSubPix().