Geometry and Critical Configurations of Multiple Views
(2001) In Doctoral Theses in Mathematical Sciences 2001:4. Abstract
 This thesis is concerned with one of the core problems in computer vision, namely to reconstruct a real world scene from several images of it. The interplay between the geometry of the scene, the cameras and the images is analyzed. The framework is based on projective geometry, which is the natural language for describing the geometry of multiple views. The affine and Euclidean geometries are regarded as special cases of projective geometry.
First, the projection of several different geometric primitives in multiple views is described and analyzed. The analysis includes points, lines, quadrics and curved surfaces. The cameras are assumed to be uncalibrated and both the perspective/projective and the affine camera model... (More)  This thesis is concerned with one of the core problems in computer vision, namely to reconstruct a real world scene from several images of it. The interplay between the geometry of the scene, the cameras and the images is analyzed. The framework is based on projective geometry, which is the natural language for describing the geometry of multiple views. The affine and Euclidean geometries are regarded as special cases of projective geometry.
First, the projection of several different geometric primitives in multiple views is described and analyzed. The analysis includes points, lines, quadrics and curved surfaces. The cameras are assumed to be uncalibrated and both the perspective/projective and the affine camera model are considered. Several new reconstruction methods are developed. Some features of these methods include the possibility of handling: (i) missing data, (ii) several different primitives simultaneously and (iii) minimal cases.
Then, focus is turned to the process of obtaining a Euclidean reconstruction of the scene from uncalibrated images. This problem is known as autocalibration. A reconstruction method which imposes regularity constraints on the camera motion is introduced which makes the autocalibration problem more stable.
The last part of the thesis is devoted to a theoretical study of necessary and sufficient conditions for obtaining a unique scene reconstruction. One classical result is that for two images of a 3D scene this happens if and only if the scene points and the camera centres do not lie on a ruled quadric. This is generalized to any number of views. Furthermore, analogous critical configurations for the 1D camera are derived. In autocalibration, it is shown that the critical configurations depend only on the camera motion. Complete classifications of such critical motions which lead to ambiguous reconstructions are given under different settings. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/20228
 author
 Kahl, Fredrik ^{LU}
 supervisor
 opponent

 Professor Zisserman, Andrew, Oxford University
 organization
 publishing date
 2001
 type
 Thesis
 publication status
 published
 subject
 keywords
 algebra, algebraic geometry, field theory, Number Theory, Matematik, Mathematics, reconstruction, image sequence, absolute conic, critical motions, critical surfaces, perspective projection, affine geometry, Euclidean geometry, multiple view geometry, projective geometry, group theory, Talteori, fältteori, algebraisk geometri, gruppteori, Mathematical logic, set theory, combinatories, Matematisk logik, mängdlära, kombinatorik
 in
 Doctoral Theses in Mathematical Sciences
 volume
 2001:4
 pages
 172 pages
 publisher
 Fredrik Kahl, Centre for Mathematical Sciences, P.O. Box 118, SE221 00 Lund, Sweden,
 defense location
 MH:C, Matematikcentrum
 defense date
 20010921 13:15:00
 ISSN
 14040034
 ISBN
 9162849352
 language
 English
 LU publication?
 yes
 id
 7813c2c348e54b0db0c3f3ddf977624c (old id 20228)
 date added to LUP
 20160401 15:17:41
 date last changed
 20190521 13:28:49
@phdthesis{7813c2c348e54b0db0c3f3ddf977624c, abstract = {This thesis is concerned with one of the core problems in computer vision, namely to reconstruct a real world scene from several images of it. The interplay between the geometry of the scene, the cameras and the images is analyzed. The framework is based on projective geometry, which is the natural language for describing the geometry of multiple views. The affine and Euclidean geometries are regarded as special cases of projective geometry.<br/><br> <br/><br> First, the projection of several different geometric primitives in multiple views is described and analyzed. The analysis includes points, lines, quadrics and curved surfaces. The cameras are assumed to be uncalibrated and both the perspective/projective and the affine camera model are considered. Several new reconstruction methods are developed. Some features of these methods include the possibility of handling: (i) missing data, (ii) several different primitives simultaneously and (iii) minimal cases.<br/><br> <br/><br> Then, focus is turned to the process of obtaining a Euclidean reconstruction of the scene from uncalibrated images. This problem is known as autocalibration. A reconstruction method which imposes regularity constraints on the camera motion is introduced which makes the autocalibration problem more stable.<br/><br> <br/><br> The last part of the thesis is devoted to a theoretical study of necessary and sufficient conditions for obtaining a unique scene reconstruction. One classical result is that for two images of a 3D scene this happens if and only if the scene points and the camera centres do not lie on a ruled quadric. This is generalized to any number of views. Furthermore, analogous critical configurations for the 1D camera are derived. In autocalibration, it is shown that the critical configurations depend only on the camera motion. Complete classifications of such critical motions which lead to ambiguous reconstructions are given under different settings.}, author = {Kahl, Fredrik}, isbn = {9162849352}, issn = {14040034}, language = {eng}, publisher = {Fredrik Kahl, Centre for Mathematical Sciences, P.O. Box 118, SE221 00 Lund, Sweden,}, school = {Lund University}, series = {Doctoral Theses in Mathematical Sciences}, title = {Geometry and Critical Configurations of Multiple Views}, volume = {2001:4}, year = {2001}, }