2008. 10. 27. 03:48
@GSMC/이상욱: Principles of Human and Artificial Vision
cross ratio
http://en.wikipedia.org/wiki/Cross_ratio
http://www.mathpages.com/home/kmath543/kmath543.htm
http://www.cut-the-knot.org/pythagoras/Cross-Ratio.shtml
measuring height: algebraic derivation from the cross ratio
Planar Perspective Map (homography) H
H Maps reference plane X-Y coordinates to image plane x-y coordinates
Fully determined from 4 known points on ground plane
Option A: physically measure 4 points on ground
Option B: find a square, guess the size
Option C: Note H = [avX bvY l] (columns 1,2,4 of P )
Play with scale factors a and b until the model “looks right”
Given x-y , can find X-Y by H-1 :
H Maps reference plane X-Y coordinates to image plane x-y coordinates
Fully determined from 4 known points on ground plane
Option A: physically measure 4 points on ground
Option B: find a square, guess the size
Option C: Note H = [avX bvY l] (columns 1,2,4 of P )
Play with scale factors a and b until the model “looks right”
Given x-y , can find X-Y by H-1 :
To compute camera projection matrix, we can measure
- Positions within a plane
- Height (More generally — distance between any two parallel planes)
- Camera position
- Positions within a plane
- Height (More generally — distance between any two parallel planes)
- Camera position
Single View Metrology
Department of Engineering Science, University of Oxford
> Assumptions
1) 3 orthogonal sets of parallel lines
2) 4 known points on ground plane
3) 1 height in the scene
> Complete approach
- Load in an image
- Click on parallel lines defining X, Y, and Z directions
- Compute vanishing points
- Specify points on reference plane, ref. height
- Compute 3D positions of several points
- Create a 3D model from these points
- Extract texture maps
- Output a VRML model
1) 3 orthogonal sets of parallel lines
2) 4 known points on ground plane
3) 1 height in the scene
> Complete approach
- Load in an image
- Click on parallel lines defining X, Y, and Z directions
- Compute vanishing points
- Specify points on reference plane, ref. height
- Compute 3D positions of several points
- Create a 3D model from these points
- Extract texture maps
- Output a VRML model
http://en.wikipedia.org/wiki/VRML
VRML (Virtual Reality Modeling Language, pronounced vermal or by its initials, originally — before 1995 — known as the Virtual Reality Markup Language) is a standard file format for representing 3-dimensional (3D) interactive vector graphics, designed particularly with the World Wide Web in mind. It has been superseded by X3D.
3-D Metrology
http://research.microsoft.com/~antcrim/
single view reconstruction
http://en.wikipedia.org/wiki/Ground_truth
radial distortion
http://en.wikipedia.org/wiki/Distortion_(optics)
'@GSMC > 이상욱: Principles of Human and Artificial Vision' 카테고리의 다른 글
Lecture 7: Image Filtering & Edge Detection (0) | 2008.10.28 |
---|---|
Lecture 6: Multiple-View Geometry: Stereo, Motion, Features (0) | 2008.10.27 |
Lecture 4: Projective Geometry (0) | 2008.10.26 |
Lecture 3: Linear Algebra for Computer Vision (0) | 2008.10.26 |
Lecture 2: Image Formation & Camera Calibration (0) | 2008.10.24 |