ARKit mapping from screen coordinates to Camera frame coordinates. Measurments application

Im working on the following problem:

For a measurment application i want to take a picture of something laying on the ground, and given i will have the floor plane detecred, i plan to raycast 4 points from the corners of the screens, given the raycast land on this plane, i want to use those coordinates to do a perspective transform (warp) of the camera-image onto the new coordinates. This way i should be able to perform pixel-per-cm measurments. The prolem i have is the screen coordinates dont seem to reflect the camera-frame coordinates, and im not sure how to go from one to another.