A Continuous Probabilistic Scene Model for Aerial Imagery

Crispell, Daniel Elting

Full Metadata

Overview

Year:: 2009
Contributor:: Crispell, Daniel Elting (creator); Taubin, Gabriel (director); Mundy, Joseph (reader); Kimia, Benjamin (reader); Brown University. Division of Engineering. Electrical Sciences and Computer Engineering (sponsor)
Genre:: theses
Subject:: 3-d modeling; Computer vision; Computer graphics; Three-dimensional modeling
Extent:: xx, 123 p.
DOI: https://doi.org/10.7301/Z0NV9GHH

Files

Description

Abstract:: Given a set of images of a scene, it is possible to estimate both the 3-d position of points in the scene and images from viewpoints not present in the original data. These estimates cannot in general be exact, however, due to uncertainties and ambiguities present in the data. For this reason, probabilistic scene models and reconstruction algorithms are ideal due to their inherent ability to represent such uncertainties and ambiguities. Unfortunately, existing probabilistic reconstruction algorithms do not scale well to large and complex scenes, primarily due to their dependence on large three-dimensional voxel arrays. The work presented in this thesis generalizes previous probabilistic models in such a way that multiple orders of magnitude savings in storage are possible, making accurate 3-d point localization and high resolution novel view generation of large-scale outdoor scenes possible. Specifically, the inherent dependence on a discrete array of uniformly sized voxels is removed through the derivation of a continuous probabilistic model which represents uncertain geometry as a density field, allowing implementations to efficiently sample the volume in a non-uniform fashion. In addition, multiple reconstruction algorithms are presented to accommodate differing modes of operation in which imagery may be captured and used. The proposed model combined with the reconstruction and novel view generation algorithms comprise the first system capable of automatically generating photo-realistic renderings of large and complex scenes from arbitrary viewpoints based on aerial image data alone.
Notes:: Thesis (Ph.D.) -- Brown University (2010)

Content

Access Conditions

Rights: In Copyright
Restrictions on Use: Collection is open for research.

Citation

Crispell, Daniel Elting, "A Continuous Probabilistic Scene Model for Aerial Imagery" (2009). Engineering Theses and Dissertations, Electrical Sciences and Computer Engineering Theses and Dissertations. Brown Digital Repository. Brown University Library. https://doi.org/10.7301/Z0NV9GHH

Relations

Collections:

Engineering Theses and Dissertations

Theses and Dissertations for the Engineering department.
...
Electrical Sciences and Computer Engineering Theses and Dissertations

Theses and Dissertations for the Electrical Sciences and Computer Engineering department.
...