Titre : |
Orientation of optical airborne and spaceborne images for small and medium scale mapping purposes |
Type de document : |
Rapport |
Auteurs : |
Eija Honkavaara, Auteur ; Mirjam Bilker, Auteur ; X. Yu, Auteur ; Juha Jaakkola, Auteur ; Harri Kaartinen, Auteur ; J. Ylonen, Auteur |
Editeur : |
Helsinki : Finnish Geodetic Institute FGI |
Année de publication : |
1999 |
Collection : |
Reports of the Finnish Geodetic Institute num. 99-5 |
Importance : |
57 p. |
Format : |
21 x 30 cm |
Note générale : |
Bibliographie |
Langues : |
Anglais (eng) |
Descripteur : |
[Vedettes matières IGN] Photogrammétrie numérique [Termes IGN] aérotriangulation [Termes IGN] image aérienne [Termes IGN] image JERS [Termes IGN] image optique [Termes IGN] image spatiale [Termes IGN] orientation d'image [Termes IGN] orientation du capteur [Termes IGN] photographie à moyenne échelle [Termes IGN] photographie à petite échelle [Termes IGN] point d'appui [Termes IGN] point de vérification [Termes IGN] points homologues
|
Index. décimale : |
33.30 Photogrammétrie numérique |
Résumé : |
(Auteur) Recent technical progress has resulted in significant improvement of the orientation process. The most revolutionary developments include automation of tie point measurement and direct determination of orientation parameters with GPS and INS techniques. Options to be taken into account when selecting materials for mapping and methods for orientation include image type (scale; standard frame, digital camera; spaceborne, airborne), block geometry (single image orientation, aerial triangulation (AT)), ground control (targeted, untargeted, GIS data, etc.), image observations (automatic, interactive) and utilisation of directly measured orientation parameters. The task is to optimise costs, accuracy and reliability.
These questions have been investigated at the Finnish Geodetic Institute (FGI) for medium and small scale mapping applications with spatial resolution and accuracy requirements for 15 m. Both theoretical and empirical considerations have been made with respect to several variables: observations of orientation parameters (GPS, no GPS, GPS processing), block structure (cross strips, no cross strips; single image, block), ground control points (GCPs, targeted, untargeted and GIS data: nationwide digital orthophotos, base maps and elevation models), image observations of tie points and GCPs (interactive, automatic) and image material (1:31 000 and 1:60 000 standard airborne images with approximately 1 m and 5 m pixel sizes; JERS/VNIR satellite images). All the airborne cases were analysed using the same strategies and checkpoints. JERSIVNIR images were analysed in parallel with the airborne images, but no comprehensive comparisons between the materials could be made because of large resolution differences.
The results of GPS measurements of perspective centres showed offsets and linear drifts varying from strip to strip, but compensation in AT with minimum ground control was possible. Targeted GCPs yielded high accuracy and reliability. Untargeted GCPs and GIS data as ground control, especially digital orthophotos and DEM25, also yielded good accuracy, although care is needed to avoid problems with gross errors. With GIS data, systematic errors can also be expected. The empirical and simulation results obtained make it possible to select GCP materials for obtaining various accuracy levels in the cases of spatial resection and AT with or without GPS support. It was concluded that in block adjustment, a planimetric accuracy of 10 gm in the image and a height accuracy of 0.1%oH can be obtained by using a reasonable number of GCPs (e.g. in 4 base distances) with accuracy better than 20 µm in image. Several examples of suitable GCP materials and configurations for obtaining accuracy better than 1 m were given. High accuracy was obtained with automatic tie point measurement, but investigations are still needed concerning optimal adjustment methods and number of points. The results, however, were in any case sufficient for applications with medium quality requirements. Experiences of orientation of JERS/VNIR images with rigorous methods were good; accuracy of 0.4 pixels was obtained with only 4 GCPs. In all the cases, the results of empirical and theoretical investigations correlated as long as the assumptions were valid; therefore reliable knowledge of the quality of the materials used as the basis for orientation is an important presupposition. |
Numéro de notice : |
67564 |
Affiliation des auteurs : |
non IGN |
Thématique : |
IMAGERIE |
Nature : |
Rapport |
Permalink : |
https://documentation.ensg.eu/index.php?lvl=notice_display&id=44680 |
| |