Why should inverse intrinsic matrix be [3,4] and not [3,3]?

[Paul45577]

Hi,

I have a problem understanding why pixtocams is [N, 3, 4]. More precisely, I don't quite understand why the inverse intrinsic matrix is a 3x4 matrix or matrices and not a 3x3 matrix.

I calculated the camera matrix with the intrinsic camera parameters (focal length, principal point) and passed it to the np.linalg.inv() function. For pixtocam I get a 3x3 matrix as return value and not a 3x4 matrix as described in the docs. Can someone please tell me what I have misunderstood here? Thanks in advance!