Using own dataset causing problems with rotation matrices

Question

Using own dataset causing problems with rotation matrices

Closed this issue 2 years ago · 0 comments

When running the pipeline on my own dataset, I get this error. Nowhere in the configuration files is this rotation matrix specified, and the slam pose transformation matrices input all use orthonomal rotation matrices. Is this a problem with my calibration somewhere?

0%| | 0/573 [00:00<?, ? frames/s] WARNING: Logging before InitGoogleLogging() is written to STDERR F20221202 16:15:29.675740 938011 rotation-quaternion-inl.h:115] Check failed: isValidRotationMatrix(matrix) -0.999833 -0.99952 -0.999484 -0.999703 -0.998442 -0.998706 -0.997548 -0.999331 -0.997702 *** Check failure stack trace: *** Aborted (core dumped)