Prediction of geometric geoid heights by artificial neural networks
Article Sidebar
Main Article Content
Abstract
Level heights of the national vertical networks (NVR), is the base for the establishment of the fundamental vertical reference datum of each country. When these heights are not associated to observations of terrestrial gravity field (purely geometric characteristics), they lose usefulness in solving problems related to regional and global physical phenomena generated due to the dynamic of the Earth System. In addition, different Local Vertical Datums (LVD) does not materialize the same reference, because the determination of mean sea level (MSL) in each LVD generates discrepancies in relation to a Global Vertical Reference System (GVRS). Geometrical heights from satellite-based positioning (GNSS), could be used for the determination of physical heights whether the relationship between the reference ellipsoid (reference surface for the determination of geometrical heights) and the reference surface based on terrestrial gravity field exists. One approximation to that relation is obtained from the difference between the ellipsoidal and level heights (hHn). The present study shows that the Artificial Neural Networks (ANN) are a suitable alternative for the prediction of geometrical heights based on this approximation. The multilayer ANN training was performed from the basic information come from records of the Ecuadorian NVR (NVRE) associated with GPS observa tions (GPS/Levelling). The results show that the prediction of heights has a maximum error of 7 cm
Downloads
Article Details
By submitting the originals, the author accepts the rules of distribution, reproduction and copyright established by the journal.
The author who publishes in Revista Cartográfica agrees with the following terms:
- It preserves the copyright, and grants the journal the right of publication (reproduction, edition, distribution, public communication and exhibition).
- Revista Cartográfica is licensed under the terms of Creative Commons Reconocimiento-NoComercial-SinObraDerivada 4.0 Internacional.
- The author is encouraged to reproduce and host their works published in the Revista Cartográfica in institutional repositories, web pages, etc. in order to contribute to the dissemination and transfer of knowledge.
- The journal adheres to the international standards and codes of the Committee on Publication Ethics (Code of Conduct and Best Practices Guidelines for Journals Editors, COPE).