Hello,
I am working as a surveyor and trying to understand covariance handling in baseline processing for comparison purposes.
From Trimble Business Center baseline processing, I obtain covariance-related outputs (either:
sdx, sdy, sdz, sdxy, sdyz, sdzx (ECEF), or sdn, sde, sdu, sdne, sdeu, sdun (ENU)
These values are reported in meters (standard deviations and covariances).
What I Am Doing
Rotating it to ECEF (XYZ) using:
Where R is the ENU→ECEF rotation matrix based on geodetic latitude and longitude.
For example, I am getting below values in TBC:
My Questions
Is this the correct approach to reconstruct the covariance matrix in ECEF (Aposteriori Covariance Matrix (Meter² in baseline)?
Should the covariance values be interpreted directly as meters² (after squaring standard deviations)?
Does RTKLIB internally:
Apply any variance factor scaling (σ₀²)?
Normalize or weight covariance differently during baseline processing?
If starting from RTKLIB .pos output (sdn, sde, sdu, sdne, sdeu, sdun), what is the recommended way to reconstruct the full 3×3 covariance matrix properly?
My goal is simply to ensure that I am constructing the covariance matrix correctly when transforming between ENU and ECEF for quality control comparison.
If there is any reference to how RTKLIB forms and scales covariance during baseline estimation, that would be very helpful.
Thank you.
Hello,
I am working as a surveyor and trying to understand covariance handling in baseline processing for comparison purposes.
From Trimble Business Center baseline processing, I obtain covariance-related outputs (either:
sdx, sdy, sdz, sdxy, sdyz, sdzx (ECEF), or sdn, sde, sdu, sdne, sdeu, sdun (ENU)
These values are reported in meters (standard deviations and covariances).
What I Am Doing
Rotating it to ECEF (XYZ) using:
Where R is the ENU→ECEF rotation matrix based on geodetic latitude and longitude.
For example, I am getting below values in TBC:
My Questions
Is this the correct approach to reconstruct the covariance matrix in ECEF (Aposteriori Covariance Matrix (Meter² in baseline)?
Should the covariance values be interpreted directly as meters² (after squaring standard deviations)?
Does RTKLIB internally:
Apply any variance factor scaling (σ₀²)?
Normalize or weight covariance differently during baseline processing?
If starting from RTKLIB .pos output (sdn, sde, sdu, sdne, sdeu, sdun), what is the recommended way to reconstruct the full 3×3 covariance matrix properly?
My goal is simply to ensure that I am constructing the covariance matrix correctly when transforming between ENU and ECEF for quality control comparison.
If there is any reference to how RTKLIB forms and scales covariance during baseline estimation, that would be very helpful.
Thank you.