realsense R200, difference between saved images and live stream?

edit

If what you want is to get an infrared image in CV_8U, you have two options for me:

• request directly an infrared image at this format: dev->enable_stream(rs::stream::infrared, 640, 480, rs::format::y8, 30);

• just convert the native CV_16U infrared image to the wanted format if you cannot change the stream for whatever reason (you need to scale to the destination range, otherwise the image is saturated):

  Mat infrared_8U;
  infrared.convertTo(infrared_8U, CV_8U, 1/256.0);
  

You can also directly shift the bits:

Mat infrared_8U(infrared.size(), CV_8U);
for (int i = 0 ;i < infrared_8U.rows; i++) {
  for (int j = 0; j < infrared_8U.cols; j++) {
    infrared_8U.at<uchar>(i,j) = infrared.at<ushort>(i,j) >> 8;
  }
}

Not sure what happens if you write a CV_16U image on disk as you seem to do.

Supported formats for each RealSense camera can be found here. Getting started for the librealsense and OpenCV: https://github.com/IntelRealSense/librealsense/blob/master/doc/stepbystep/getting_started_with_openCV.md.

Demo code tested with an Intel SR300 camera:

// Create a context object. This object owns the handles to all connected realsense devices
rs::context ctx;

// Access the first available RealSense device
rs::device * dev = ctx.get_device(0);

// Configure Infrared stream to run at VGA resolution at 30 frames per second
dev->enable_stream(rs::stream::infrared, 640, 480, rs::format::y16, 30);

// Start streaming
dev->start();

// Camera warmup - Dropped several first frames to let auto-exposure stabilize
for(int i = 0; i < 30; i++) {
  dev->wait_for_frames();
}

// Creating OpenCV Matrix from a color image
Mat infrared(Size(640, 480), CV_16U, (void*)dev->get_frame_data(rs::stream::infrared), Mat::AUTO_STEP);

Mat infrared_8U;
infrared.convertTo(infrared_8U, CV_8U, 1/256.0);

// Display in a GUI
namedWindow("Display Image", WINDOW_AUTOSIZE );
imshow("Display Image", infrared_8U);

waitKey(0);