I am a complete beginner to 3D Reconstruction and Epipolar geometry.

I've set up a pair of stereo cameras (independent cameras mounted on separate axis). These cameras have been synchronized and I have calibrated the stereo pair using camera_calibration "cameracalibrator.py".

My goal is to simply extract a 3D world pose (X,Y,Z) (in the form of a tf marker) given a 2D image coordinate (u,v) on one of the cameras. (Essentially the opposite of what is done in this tutorial.)

How do I utilise the generated "P" and "R" matrices and/or the image_geometry "projectPixelTo3dRay()" function to achieve this? I was unable to find any code examples of "ray-plane intersection" as described in this answer.