[rig] remove the constraint that main camera has the same reference system of the rig

[simogasp]

Currently, we assume that the pose of the rig is the same as the pose of the main camera. Let the main camera have its own subpose (possibly the [I 0]). This affects:

• [ ] rigResection
• [ ] refineRig
• [ ] all the localizeRig()
• [ ] calibration
• [ ] load/save calibration files
• [ ] all the doc

See comments in PR alicevision/AliceVision#190 as a guide

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[alexs7]

@simogasp Hello, I wanted to ask about the rigResection and the vec_subPoses variable. The definition of a subPose currently is " the subposes are always wrt the main camera". The main camera is assigned to the identity matrix. The subPoses will have to be the relative poses from the origin (the identity matrix, main camera) then ?

[simogasp]

Yes, the subPoses are always expressed wrt the main camera, so that the main camera will have as sub-pose the Identity matrix and all the others will have a subPose expressed wrt the main.

The main camera (which basically represents the whole rig) has a pose P expressed wrt the world reference. To get the absolute pose of a camera of the rig you have to multiply P by the relevant subPose.

[alexs7]

Ok thanks!