The adjustment of planned survey networks, as well as the substantiated analysis of existing geodetic networks using the principles of least squares adjustment,
is one of the most important tasks in metrology and the geodetic sciences, such as surveying engineering. In addition to the adjusted coordinates and the corresponding
fully populated dispersion matrix, the network adjustment provides essential parameters for evaluating the so-called internal and external reliability of the network.
Nowadays, software-based analysis of geodetic networks is state of the art and not reserved for special applications or a small group of experts. Just as the experienced
handling of the measuring equipment and instrumentation is a necessary condition for precise data, a proper use of the application is essential to obtain reliable results.
This is the only way to correctly interpret adjustment results and to achieve an additional benefit in further analysis procedures.
Quality assurance through round robin tests
In order to guarantee traceable results, but also to ensure comparability with other software packages, scientifically supervised round robin tests are essential.
In general, the reasons for discrepancies in adjustment results are complex, e. g., hidden or undocumented preprocessing steps, but also different implementations of
the functional as well as the stochastic model.
Validation of algorithms implemented in JAG3D for best fit of form elements by the Physikalisch-Technische Bundesanstalt (PTB), cf. full report as pdf
Since Java·Applied·Geodesy·3D is a freely available open-source software package, universities and metrological institutes often study
the software and evaluate the adjustment results. JAG3D's adjustment results show a high degree of consistency with results from other network adjustment programs and
are almost identical with results obtained from well-known proprietary software packages. This could be demonstrated for small precise engineering networks as well as
for large surveying networks with an extension of several kilometres.
Data sets provided by the National Institute of Standards and Technology (NIST) were used to validate the adjustment results of estimated curves and surfaces.
According to ISO 10360-6, which deals with the estimation of errors in computing Gaussian associated features, an independent verification of the results was
performed by the Physikalisch-Technische Bundesanstalt (PTB). The PTB is the National Metrology Institute and the supreme technical authority of the Federal Republic of Germany for metrology.
The entire test consisted of 44 sets of data describing basic features such as straight line, plane, circle, cylinder, sphere and cone. The algorithms implemented in JAG3D were verified using these sets.
The resulting maximum deviations (MPE) of the estimates were less than 0.1 μm and 0.1 μrad and confirm the obtained adjustment results,
as shown in the full TraCIM test report of the PTB.
Quality assurance through training
Tutorial analysis of geodetic networks using JAG3D at 20th International Course on Engineering Surveying at ETH Zurich
As part of the International Courses on Engineering Surveying at ETH Zurich and Graz University of Technology, the fundamentals of geodetic network adjustment were recapitulated through specific, practice-oriented exercises.
Under expert guidance, participants used the open source least-squares software package JAG3D to analyse local terrestrial networks, to interpret the adjustment results, and to
evaluate network reliability. Particular attention was also paid to the limitations of the least-squares adjsutment, which often lead to misinterpretations of the results and are frequently omitted
from training sessions.
The principles of Baarda's DIA approach for detection, identification and adaptation of misspecified models were generalised and applied to evaluate network
deformations. In complex analysis procedures, the network adjustment is only a cog in the wheel. For that reason, interfaces for data exchange and processing
of adjustment results in external applications were addressed.
Please contact us if you are interested in a refresher course or further qualification for yourself or your team. Learn to value what matters.
References
Lösler, M., Bähr, H.: Vergleich der Ergebnisse verschiedener Netzausgleichungsprogramme. In: Zippelt, K. (ed.): Vernetzt und ausgeglichen - Festschrift zur Verabschiedung von Prof. Dr.-Ing. habil. Dr.-Ing. E.h. Günter Schmitt, KIT Scientific Publishing, 205-214, 2010. DOI: 10.5445/KSP/1000020074
Durand, S., Lösler, M., Jones, M., Cattin, P.-H., Guillaume, S., Morel, L.: Quantification of the dependence of the results on several network adjustment applications. In: García-Asenjo, L., Lerma, J. L. (eds.): 5th Joint International Symposium on Deformation Monitoring (JISDM), Editorial Universitat Politècnica de València, 69-77, 2022. 10.4995/JISDM2022.2022.13671
Lösler, M., Eschelbach, C.: Netzanalyse: Beurteilung von Ausgleichungsergebnissen geodätischer Netze und deren Weiterverarbeitung. 20th International Course on Engineering Surveying, ETH Zurich, 2023.
Lösler, M., Eschelbach, C., Kopitzke K.-L.: Softwarebasierte Ausgleichung geodätischer Netze: Analyse, Interpretation und Weiterverarbeitung. 21th International Course on Engineering Surveying, TU Graz, 2026.
