What does this piece of code do in this OpenCV4Android sample application?

Question

Here is the ColorBlobDetectionActivity class of the color blob detection sample, and the particular chunk of code that I am facing difficulty in understanding is Line#114 to Line # 135 in the onTouch method implemented in this class.

When the onTouch method is invoked, that is when the user touches a colored blob, the int rows= mRgba.rows() and int cols = mRgba.cols() is calculated. Since mRgba is a Mat which was returned by onCameraFrame(), it means it represents a camera frame.So I think rows and cols now represent the number of pixels along x-axis, and that along y-axis of the frame.

Since a frame is the area viewed by the camera (which in this app is the full screen of the device), so rows and cols represent the number of pixels along x-axis and y-axis of the screen respectively.

The next two statements are:

int xOffset = (mOpenCvCameraView.getWidth() - cols) / 2;
int yOffset = (mOpenCvCameraView.getHeight() - rows) / 2;

The questions are:

1. What exactly do xOffset and yOffset represent?

2. mOpenCvCameraView is an instance of CameraBridgeViewBase, which according to the documentation is a basic class responsible for implementing the interaction of Camera and OpenCV. The documentation on getWidth() and getHeight() is silent, but I think it is also the width and height of the camera frame (in pixels?), so it should be same as rows and cols. Is that correct?

3. Can you explain a bit the formula they have used to calculate xOffset and yOffset (in the above two statemnets)?

Answer 1

The delivered frames and the surface size are not necessarily the same. The maximum of the event.getX equals to the surface width.

I haven't seen the program running, but it seems like the offset determines the size of the touched rect. The rect is used for averaging the color, not simply displaying a single pixel's data.