The current method of aligning cameras depends on using the built-in Blender camera solver and subsequently aligning the solve based on identical tracks. This means that even if the second camera is intended to be static, there needs to be initial non-nodal (hand-held) movement to provide perspective parallax for the solve.
Open CV
cv2.solvePnP() seems to be able to provide a 3D camera position based on a set of existing coordinates and their corresponding image UV given a camera matrix.
If a single static camera can be oriented in 3D space it stands to reason that multiple static cameras could be properly oriented. Continuing down that rabbit hole, this could turn into an application for an entire motion capture studio.
Situation as it Stands
The current method of aligning cameras depends on using the built-in Blender camera solver and subsequently aligning the solve based on identical tracks. This means that even if the second camera is intended to be static, there needs to be initial non-nodal (hand-held) movement to provide perspective parallax for the solve.
Open CV
cv2.solvePnP()seems to be able to provide a 3D camera position based on a set of existing coordinates and their corresponding image UV given a camera matrix.OpenCV Documentation
Explanation of Perspective-n-Point
Going Further
If a single static camera can be oriented in 3D space it stands to reason that multiple static cameras could be properly oriented. Continuing down that rabbit hole, this could turn into an application for an entire motion capture studio.