I gone through some similar post for this problem like this,
@RafazZ
OpenCV unproject 2D points to 3D with known depth `Z`
2D Coordinate to 3D world coordinate
as doing 2d to 3d conversion in camera_frame they have removed R,T matrix from calculation and only using like this,
x = (u-cx)/fx
y= (v-cy)/fy
actual (X,Y,Z)
X= x * depth.at(u,v)
Y= y * depth.at(u,v)
Z = depth.at(u,v)
so, I want to understand why we can remove R,t from calculation like this when doing 2d-->3d in camera_frame?