Camera Settings

During image alignment or registration process, RealityCapture calculates parameters and the pose where was the image taken. This camera pose is illustrated in the 3D view as a small white cone with a small image.

Providing more information about cameras which were used can simplify and speed-up this process.

Calibration Groups

By defining a calibration group we state that all images in this group have the same properties, e.g. the same focal length, the same principal point or the same lens distortion coefficients. These groups can be created automatically using image metadata, a user can alternatively create groups manually.

What are calibration groups good for?

To give you a use-case example, using calibration groups can be useful:

In general, images with the same body/optics must have the same or very similar calibration parameters. These parameters may vary due to small movements of optics caused by focusing, mechanical movement or temperature changes. These differences are negligible for small resolution cameras, while for high resolution cameras, e.g. 60 Mpx and more, these differences are more visible. On the other hand, by grouping calibration parameters together RealityCapture will need to calculate less parameters which results in a faster computation.

Automatic Grouping

Manual Grouping

There is also a possibility to group images individually.

TIP: After calculating a scene with grouped camera parameters, ungroup camera parameters and align the scene again. It will fine-tune the camera parameters and the scene also stays well-conditioned. This is recommended for scenes with weak texture or small image count. Alignment in such case will take just a few seconds.

Camera Priors

Define lens distortion priors

See also: