Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more
Standard view
Full view
of .

1.1 Visual perception: from 2-D images to 3-D models
1.2. A historical perspective 3
1.2 A historical perspective
1.3 A mathematical approach
1.4. Organization of the book 5
1.4 Organization of the book
Introductory material
Representation of a three dimensional moving scene
2.1 Three-dimensional Euclidean space
2.2 Rigid body motion
2.3 Rotational motion and its representations
2.3.1 Canonical exponential coordinates
2.3.2 Quaternions and Lie-Cartan coordinates
2.4 Rigid body motion and its representations
2.4.1 Homogeneous representation
2.4.2 Canonical exponential coordinates
2.5 Coordinates and velocity transformation
2.6 Summary
2.7 References
2.8 Exercises
Image formation
3.1 Representation of images
3.2 Lenses, surfaces and light
3.2.1 Imaging through lenses
3.2.2 Imaging through pin-hole
3.3. A ﬁrst model of image formation 43
3.3 A ﬁrst model of image formation
3.3.1 Basic photometry
3.3.2 The basic model of imaging geometry
3.3.3 Ideal camera
3.3.4 Camera with intrinsic parameters
3.3.5 Spherical projection
3.3.6 Approximate camera models
3.4 Summary
3.5 Exercises
Image primitives and correspondence
4.1 Correspondence between images
4.1.1 Transformations in the image domain
4.1.2 Transformations of the intensity value
4.2 Photometric and geometric features
4.3 Optical ﬂow and feature tracking
4.3.1 Translational model
4.3.2 Aﬃne deformation model
4.4 Feature detection algorithms
4.4.2 Line features: edges
4.4.3 Point features: corners
4.6 Exercises
Reconstruction from two calibrated views
5.1 The epipolar constraint
5.1.1 Discrete epipolar constraint and the essential matrix
5.1.2 Elementary properties of the essential matrix
5.2 Closed-form reconstruction
5.2.1 The eight-point linear algorithm
5.2.2 Euclidean constraints and structure reconstruction
5.3 Optimal reconstruction
5.4 Continuous case
5.4.2 Properties of continuous essential matrices
5.4.3 The eight-point linear algorithm
5.4.4 Euclidean constraints and structure reconstruction
5.5 Summary
5.6 Exercises
Camera calibration and self-calibration
6.1 Calibration with a rig
6.2 The fundamental matrix
6.2.2 The eight-point linear algorithm
6.3 Basics of uncalibrated geometry
6.3.1 Kruppa’s equations
6.4. Self-calibration from special motions and chirality 127
6.4 Self-calibration from special motions and chirality
6.4.1 Pure rotational motion
6.4.2 Translation perpendicular or parallel to rotation
6.4.3 Calibration with chirality
6.5 Calibration from continuous motion
6.6 Three stage stratiﬁcation
6.6.1 Projective reconstruction
6.6.2 Aﬃne reconstruction
6.6.3 Euclidean reconstruction
6.7 Summary
6.8 Exercises
Introduction to multiple view reconstruction
7.1 Basic notions: pre-image and co-image of point and line
7.2 Preliminaries
7.3. Pairwise view geometry revisited 159
7.3 Pairwise view geometry revisited
7.4. Triple-wise view geometry 161
7.4 Triple-wise view geometry
7.5 Summary
7.6 Exercises
Geometry and reconstruction from point features
8.1 Multiple views of a point
8.2 The multiple view matrix and its rank
8.3. Geometric interpretation of the rank condition 173
8.3 Geometric interpretation of the rank condition
8.3.1 Uniqueness of the pre-image
8.4.2 Reconstruction
8.5 Experiments
8.5.1 Setup
8.5.2 Comparison with the 8 point algorithm
8.5.3 Error as a function of the number of frames
8.5.4 Experiments on real images
8.6 Summary
8.7 Exercises
Geometry and reconstruction from line features
9.1 Multiple views of a line
9.2. The multiple view matrix and its rank 189
9.2 The multiple view matrix and its rank
9.3. Geometric interpretation of the rank condition 191
9.3 Geometric interpretation of the rank condition
9.3.1 Uniqueness of the pre-image
9.3.2 Geometry of the multiple view matrix
9.4. Applications of the rank condition 197
9.4 Applications of the rank condition
9.4.1 Matching
9.4.2 Reconstruction
9.5 Experiments
9.5.1 Setup
9.5.2 Motion and structure from four frames
9.5.3 Error as a function of number of frames
9.6 Summary
Geometry and reconstruction with incidence relations
10.1 Image and co-image of a point and line
10.2 Rank conditions for various incidence relations
10.2.1 Inclusion of features
10.2.2 Intersection of features
10.2.3 Features restricted to a plane
10.3 Rank conditions on the universal multiple view matrix
10.4 Geometric interpretation of the rank conditions
10.5. Applications of the rank conditions 225
10.5 Applications of the rank conditions
10.6 Simulation results
10.7 Summary
10.8 Exercises
Multiple view geometry in high dimensional space
11.1 Projection in high dimensional space
11.1.1 Image formation
11.1.2 Motion of the camera and homogeneous coordinates
11.1.3 Preimage and co-image
11.1.4 Generic motion assumption
11.2 Rank condition on multiple images of a point
11.3.2 Inclusion of hyperplanes
11.3.3 Intersection of hyperplanes
11.3.4 Restriction to a hyperplane
11.4. Geometric interpretation of rank conditions 251
11.4 Geometric interpretation of rank conditions
11.4.1 Multiple view stereopsis in n-dimensional space
11.4.2 A special case
11.4.3 Degenerate motions
11.5 Applications of the generalized rank conditions
11.5.1 Multiple view constraints for dynamical scenes
11.5.2 Multiple views of a kinematic chain
11.6 Summary
Reconstruction algorithms
Batch reconstruction from multiple views
12.1 Algorithms
12.1.1 Optimal methods
12.1.2 Factorization methods
Recursive estimation from image sequences
13.1 Motion and shape estimation as a ﬁltering problem
13.2 Observability
13.3 Realization
13.4 Stability
13.5 Implementation
13.6 Complete algorithm
Step-by-step building of a 3-D model from images
14.1 Establishing point correspondence
14.1.1 Feature extraction
14.1.2 Feature matching
14.2 Reﬁning correspondence
14.2.1 Computing fundamental matrices
14.2.2 Robust matching via RANSAC
14.3 Uncalibrated 3-D structure and camera pose
14.3.1 Projective reconstruction
14.4 Scene and camera calibration
14.4.1 Constraints on the scene
14.4.2 Camera self-calibration
14.5 Dense reconstruction
14.5.1 Epipolar rectiﬁcation
14.5.2 Dense matching
14.5.3 Knowledge in the scene
14.5.4 Multi-scale matching
14.5.5 Dense triangulation for multiple frames
14.6 Texture mapping and rendering
14.6.1 Modeling surface
14.6.2 Highlights and specular reﬂections
Extensions, applications and further research directions
15.1 Vision for navigation and robotic control
15.2 Multiple linked rigid bodies and motion segmentation
15.3 Beyond geometric features
15.4 Direct methods
15.5 Non-lambertian photometry
Appendices
Basic facts from linear algebra
A.1 Linear maps and linear groups
A.2 Gram-Schmidt orthonormalization
A.3 Symmetric matrices
A.4 Structure induced by a linear map
A.5. The Singular Value Decomposition (SVD) 299
A.5 The Singular Value Decomposition (SVD)
A.5.1 Algebraic derivation
A.5.2 Geometric interpretation
A.5.3 Some properties of the SVD
Least-square estimation and ﬁltering
B.1 Linear least-variance estimators of random vectors
B.1.1 Projections onto the range of a random vector
B.1.2 Solution for the linear (scalar) estimator
B.1.3 Aﬃne least-variance estimator
B.2 Linear least-variance estimator for stationary processes
B.3 Linear, ﬁnite-dimensional stochastic processes
B.4 Stationariety of LFDSP
0 of .
Results for:
P. 1
IP-3dvision

# IP-3dvision

Ratings: (0)|Views: 2,995|Likes:

### Availability:

See more
See less

11/03/2011

pdf

text

original

Pages 7 to 85 are not shown in this preview.