Incorrect 3D placement of 'smarter' rigid bodies

[Armita Manafzadeh]

When 'smart' rigid bodies are calculated using only 1-2 markers in one of the cameras, more than one 3D configuration is possible.

XMALab often selects the wrong one, creating jumpy rigid bodies. This generates false data that are included when rigid body transformations are exported. I have had to go in and manually delete data from tracked points, or convert .csv cells into NaNs, so that I can later run my analyses in Maya.

Can we create a way to turn off this 'smart' option rather than just deleting data? Or, preferably, can we think of a way we can get XMALab to select the correct 3D rigid body placement instead?

[Benjamin Knorlein]

fixes #89: Computation of RB pose with less than 3 2d correspondences can be disabled in the settingsdialog

→ <<cset c20e1ef4e243>>

[Armita Manafzadeh]

Ben, the dataset you asked for is attached here.