Hybrid Adjustment - heights - Java·Applied·Geodesy·3D

Greetings Micha,

I would like to do a hybrid adjustment in JAG3D combining Total station observation od 3D traverse, Leveling data and GNSS vectors. Now, GNSS vectors are derived from TBC, and proccesed in WGS84 coordinate system. I was thinking of using the WGS84 system for the whole adjustment, and then transforming the resulting coordinates into my desired system. Firstly, how does the height data derived from total station observation compare to leveling data in the software? I presume the weight of the leveling data will outweight the terestrial observations so that the heights will be determened with the leveling. Secondly, how will the leveling data combine with the adjustment? I would have a reference point with the height of the local site (my height system that I desire to have my final heights in) but the coordinates and GNSS vectors are in another system as specified before (WGS84). Can this adjustment compute? I want to leverage all the observations that I have and the end result to be the coordinates in my system including the height. Is it feaseble to transform my known reference point height to WGS84 and compute it?

Thank you in advance,
Best regards.

Hello,

I would like to do a hybrid adjustment in JAG3D combining Total station observation od 3D traverse, Leveling data and GNSS vectors. Now, GNSS vectors are derived from TBC, and proccesed in WGS84 coordinate system. I was thinking of using the WGS84 system for the whole adjustment, and then transforming the resulting coordinates into my desired system.

What is the desired system, and is it not possible to use this system to perform the adjustment process?

Firstly, how does the height data derived from total station observation compare to leveling data in the software?

Depending on the network design and dimensions, the data from the total station can be used directly. If the data refers to a real spatial network, it is not necessary to divide the raw observations (directions, zenith angles, slope distances) it into a horizontal part (direction and horizontal distances) and a height part (height differences). Moreover, each observation has its own individual uncertainty, which is used to derive a specific weight within the adjustment procedure. It is up to you and your specific weights to increase the impact of the leveling data w.r.t. the total station data.

Secondly, how will the leveling data combine with the adjustment? I would have a reference point with the height of the local site (my height system that I desire to have my final heights in) but the coordinates and GNSS vectors are in another system as specified before (WGS84). Can this adjustment compute?

Perhaps, but usually not, because you have to define a uniform datum, and - as far as I understand your explanation - this requirement is probably not fulfilled in your case.

Kind regards
Micha

--
applied-geodesy.org - OpenSource Least-Squares Adjustment Software for Geodetic Sciences

What is the desired system, and is it not possible to use this system to perform the adjustment process?

The desired system is an Mercator projection. I am not sure I can combine 3D GNSS baselines from WGS84 datum and projection points seeing as the baselines are refering to geocentric coordinate system? I thought I would atleast need to do transformation on the baselines into the system, but I still have the problem with the height. Seeing as the height is not referenced to the elipsoid but rather local orthometric. I have red recently on this matter if I recall correctly that you suggested to only use 2D baselines for such adjustment?

Perhaps, but usually not, because you have to define a uniform datum, and - as far as I understand your explanation - this requirement is probably not fulfilled in your case.

I was thinking of maybe transforming the reference point height into the same WGS84 system as the GNSS baselines and after adjusment to transform it back to my projection with the heights. Ofcourse, the height would be considered elipsoidal but that shouldn't make a difference in the end result?

Best regards.

Hello,

The desired system is an Mercator projection.

The Mercator projection is a map projection, without a defined height component. This allows you to divide your problem into two independent adjustment tasks. Separate your observations into planar observations such as GNSS-2D baselines, which should be projected into the target datum, direction and horizontal distances and adjust the horizontal network. For the height component, adjust a 1D network using the leveling data (and the height differences derived from the zenith angle and the slope distance, and - if required - the z-components of the baseline).

Seeing as the height is not referenced to the ellipsoid but rather local orthometric.

That's right. However, the ellipsoidal height is approximately the sum of the orthometric height and the geoid height, i.e. h = H + N. The difference is

Δh = hA - hb = HA + NA - HB - NB ≈ HA - HB,

if NA ≈ NB is assumed, which may valid in small networks. If NA differs significantly from NB, you must correct the geoid heights to combine these heights with your orthometric heights.

Kind regards
Micha

--
applied-geodesy.org - OpenSource Least-Squares Adjustment Software for Geodetic Sciences

RSS Feed of thread