Multiple View Geometry in Computer Vision
Multiple View Geometry in Computer Vision
Automatic Panoramic Image Stitching using Invariant Features
International Journal of Computer Vision
ICCSA'12 Proceedings of the 12th international conference on Computational Science and Its Applications - Volume Part II
EuroMed'12 Proceedings of the 4th international conference on Progress in Cultural Heritage Preservation
Hi-index | 0.00 |
This paper presents a methodology able to handle georeferenced panoramas (GeoPans) projected on 3D models for the integration of landscapes into digital environments. This is not a simple task because the typical visualization (say vertical point of view) through geographic data and GIS software does not fulfil a fundamental request: the virtual reproduction of the human eye at head height. This means that a transition from aerial images to ground (terrestrial) data is mandatory. In addition, an improvement of SDI able to generate innovative typologies of representation is needed. In this work a methodological approach aimed at rediscovering and correlating 3D reconstructions of landscapes with the typical human vision is illustrated. This contribution investigates the potential of panoramic view reconstruction and simulation from images acquired by RC/UAV and by multi-sensor terrestrial platforms (photoGPS) along with existing cartographic data. The main aim is the generation of multiple visual models able to simulate real scenarios at head height (or low altitude above ground). Examples and case studies are illustrated and discussed to prove the complexity of the problem, which requires not only new algorithms and procedures for data acquisition and processing, but also a modification of the traditional 2.5D representation of geographic data.