The camera is pre-calibrated to get the intrincs. You need to calibrate for your own camera. For each frame, the marker is detected and the RT matrix is calculated by SolvePnP. Then display the cube on the detected marker according to its estimated position.

Marker used: ArUco

For building, use Cmake