This is based on the changes in #25; merge that first.
Improved and cleaned up my absolution orientation support.
I'm now computing he quaternion covariances by first converting the input orientation into euler angles, and then computing the jacobian of the euler->quaternion transformation to convert the euler covariances into quaternion covariances.
I've also added a few tests which validate that the transform produces non-infinite results.
This is based on the changes in #25; merge that first.
Improved and cleaned up my absolution orientation support.
I'm now computing he quaternion covariances by first converting the input orientation into euler angles, and then computing the jacobian of the euler->quaternion transformation to convert the euler covariances into quaternion covariances.
I've also added a few tests which validate that the transform produces non-infinite results.