Dimension for depth in Equation 1 in csCalibration (pzV vs pxV)

[adwiii]

I am trying derive a point cloud using the equations in csCalibration.pdf, specifically equation 1:

The left-most side has pzV as the scalar factor. From my understanding of the camera projection equations (reference 1, reference 2), this is derived from the equation: screen_xy = physical_xy * focal_length / distance, which has been re-formed by moving distance to the left side as distance * screen_xy = physical_xy * focal length. From this, pzV is standing in as the distance or depth from the camera (though there are transforms to account for offset from the camera, etc).

However, the vehicle coordinate frame uses ISO 8855 conventions and the Z dimension is up instead of the standard camera coordinate system where Z represents depth. From this, it appears that the equation should use pxV instead, and when making that substitution I achieve reasonable results. This may also be related to #28.

Is this interpretation correct? Should equation 1 have pxV on the left side?