Structure from Motion ECE 847: Digital Image Processing

Slides:



Advertisements
Similar presentations
Scene Reconstruction from Two Projections Eight Points Algorithm Speaker: Junwen WU Course:CSE291 Learning and Vision Seminar Date: 11/13/2001.
Advertisements

Structure from motion.
The fundamental matrix F
Introduction to Computer Vision 3D Vision Lecture 4 Calibration CSc80000 Section 2 Spring 2005 Professor Zhigang Zhu, Rm 4439
3D Reconstruction – Factorization Method Seong-Wook Joo KG-VISA 3/10/2004.
Two-View Geometry CS Sastry and Yang
Self-calibration.
Two-view geometry.
Slides by Olga Sorkine, Tel Aviv University. 2 The plan today Singular Value Decomposition  Basic intuition  Formal definition  Applications.
Camera calibration and epipolar geometry
Vision, Video And Virtual Reality 3D Vision Lecture 13 Calibration CSC 59866CD Fall 2004 Zhigang Zhu, NAC 8/203A
Structure from motion.
Algorithms & Applications in Computer Vision
Stanford CS223B Computer Vision, Winter 2005 Lecture 11: Structure From Motion 2 Sebastian Thrun, Stanford Rick Szeliski, Microsoft Hendrik Dahlkamp and.
3D Geometry for Computer Graphics
Some useful linear algebra. Linearly independent vectors span(V): span of vector space V is all linear combinations of vectors v i, i.e.
Stanford CS223B Computer Vision, Winter 2007 Lecture 8 Structure From Motion Professors Sebastian Thrun and Jana Košecká CAs: Vaibhav Vaish and David Stavens.
Epipolar geometry. (i)Correspondence geometry: Given an image point x in the first view, how does this constrain the position of the corresponding point.
Structure from motion. Multiple-view geometry questions Scene geometry (structure): Given 2D point matches in two or more images, where are the corresponding.
Uncalibrated Geometry & Stratification Sastry and Yang
Stanford CS223B Computer Vision, Winter 2006 Lecture 8 Structure From Motion Professor Sebastian Thrun CAs: Dan Maynes-Aminzade, Mitul Saha, Greg Corrado.
Multiple-view Reconstruction from Points and Lines
3D reconstruction of cameras and structure x i = PX i x’ i = P’X i.
Structure From Motion Sebastian Thrun, Gary Bradski, Daniel Russakoff
© 2003 by Davi GeigerComputer Vision October 2003 L1.1 Structure-from-EgoMotion (based on notes from David Jacobs, CS-Maryland) Determining the 3-D structure.
Previously Two view geometry: epipolar geometry Stereo vision: 3D reconstruction epipolar lines Baseline O O’ epipolar plane.
3D Geometry for Computer Graphics
Computer Vision Structure from motion Marc Pollefeys COMP 256 Some slides and illustrations from J. Ponce, A. Zisserman, R. Hartley, Luc Van Gool, …
Single-view geometry Odilon Redon, Cyclops, 1914.
The Pinhole Camera Model
CSCE 641 Computer Graphics: Image-based Modeling (Cont.) Jinxiang Chai.
Camera parameters Extrinisic parameters define location and orientation of camera reference frame with respect to world frame Intrinsic parameters define.
CSCE 641 Computer Graphics: Image-based Modeling (Cont.) Jinxiang Chai.
CS 558 C OMPUTER V ISION Lecture IX: Dimensionality Reduction.
Multi-view geometry. Multi-view geometry problems Structure: Given projections of the same 3D point in two or more images, compute the 3D coordinates.
Euclidean cameras and strong (Euclidean) calibration Intrinsic and extrinsic parameters Linear least-squares methods Linear calibration Degenerate point.
Camera Geometry and Calibration Thanks to Martial Hebert.
Projective cameras Motivation Elements of Projective Geometry Projective structure from motion Planches : –
© 2005 Yusuf Akgul Gebze Institute of Technology Department of Computer Engineering Computer Vision Geometric Camera Calibration.
Homogeneous Coordinates (Projective Space) Let be a point in Euclidean space Change to homogeneous coordinates: Defined up to scale: Can go back to non-homogeneous.
Structure from Motion Computer Vision CS 143, Brown James Hays 11/18/11 Many slides adapted from Derek Hoiem, Lana Lazebnik, Silvio Saverese, Steve Seitz,
CSCE 643 Computer Vision: Structure from Motion
Multiview Geometry and Stereopsis. Inputs: two images of a scene (taken from 2 viewpoints). Output: Depth map. Inputs: multiple images of a scene. Output:
Geometry of Multiple Views
Affine Structure from Motion
Single-view geometry Odilon Redon, Cyclops, 1914.
Two-view geometry. Epipolar Plane – plane containing baseline (1D family) Epipoles = intersections of baseline with image planes = projections of the.
EECS 274 Computer Vision Affine Structure from Motion.
1 Chapter 2: Geometric Camera Models Objective: Formulate the geometrical relationships between image and scene measurements Scene: a 3-D function, g(x,y,z)
3D Computer Vision and Video Computing 3D Vision Topic 2 of Part II Calibration CSc I6716 Fall 2009 Zhigang Zhu, City College of New York
776 Computer Vision Jan-Michael Frahm & Enrique Dunn Spring 2013.
3D Computer Vision and Video Computing 3D Vision Topic 2 of Part II Calibration CSc I6716 Spring2013 Zhigang Zhu, City College of New York
Single-view geometry Odilon Redon, Cyclops, 1914.
Uncalibrated reconstruction Calibration with a rig Uncalibrated epipolar geometry Ambiguities in image formation Stratified reconstruction Autocalibration.
Structure from Motion Paul Heckbert, Nov , Image-Based Modeling and Rendering.
Digital Image Processing Additional Material : Imaging Geometry 11 September 2006 Digital Image Processing Additional Material : Imaging Geometry 11 September.
3D Computer Vision and Video Computing 3D Vision Topic 3 of Part II Calibration CSc I6716 Spring 2008 Zhigang Zhu, City College of New York
Structure from motion Multi-view geometry Affine structure from motion Projective structure from motion Planches : –
Instructor: Mircea Nicolescu Lecture 9
EECS 274 Computer Vision Projective Structure from Motion.
Reconstruction of a Scene with Multiple Linearly Moving Objects Mei Han and Takeo Kanade CISC 849.
55:148 Digital Image Processing Chapter 11 3D Vision, Geometry
Structure from Motion ECE 847: Digital Image Processing
Epipolar geometry.
Structure from motion Input: Output: (Tomasi and Kanade)
Uncalibrated Geometry & Stratification
George Mason University
Single-view geometry Odilon Redon, Cyclops, 1914.
Structure from motion Input: Output: (Tomasi and Kanade)
Presentation transcript:

Structure from Motion ECE 847: Digital Image Processing Stan Birchfield Clemson University

Acknowledgment Many slides are courtesy of others

SVD Any mxn matrix A can be decomposed as where This is the singular value decomposition (SVD) mxm mxn nxn

Tall and short matrices Tall matrix m>n, p = n = mxm mxn nxn Short matrix m<n, p = m = mxm mxn nxn

Compact version = = Tall matrix Tall matrix m>n, p = n mxm mxn nxn Short matrix Short matrix m<n, p = m = mxm mxn nxn

Compact version (cont.) Tall matrix Tall matrix m>n, p = n = mxn nxn nxn Short matrix Short matrix m<n, p = m = mxm mxm mxn

SVD reveals structure Let r be the index of the smallest non-zero singular value Then Easy to show:

Eigen / singular Singular values and singular vectors work like eigenvalues and eigenvectors: First p eigenvalues of ATA (or AAT) are squares of the singular values of A:

Condition number A is non-singular if and only if In real life, matrices are never singular. The condition number of A is If 1/C is near the machine’s precision, then A is ill-conditioned. It is dangerous to invert A.

Norms Singular values readily yield norms: Induced Euclidean norm: Frobenius norm: (Euclidean norm, treating matrix as vector)

Least squares where The set of equations is solved as or

Least squares (cont.) Minimum norm least squares solution to Ax=b, i.e., the shortest vector x that achieves is unique and is given by where pseudoinverse inverts all nonzero singular values

Homogeneous system What if b is all zeros? Then the minimum-norm solution is not interesting, b/c it will be x=0 always Instead, find unit-norm solution Solution is given by (the right singular vector associated with the smallest singular value)

Enforcing constraints Find closest matrix to A in the sense of Frobenius norm that satisfies constraints exactly: Factorize A = USVT Change S to S’ to satisfy constraints Put back together: A’ = US’VT Example: Enforce rank of A by setting small singular values to zero

Geometric interpretation of SVD

Structure from motion Structure from motion (SFM) recovers scene geometry camera motion from a sequence of images Could be called structure (or shape) and motion from video (SAMV), but nobody does this

SFM preliminaries Collect F frames of P points (with correspondence) Camera coordinate system: centered at focal point and aligned with image axes (x and y in image, positive z along optical axis) World coordinate system is coincident with first camera (arbitrary)

SFM under perspective projection pth point Perspective imaging: Equation counting: 2FP+1 equations (extra equation from scale ambiguity) 3P + 6(F-1) unknowns Required: 2FP+1 >= 3P + 6(F-1) With 2 frames, need at least 5 points xp-tf xp if fth camera coord sys. tf world coord sys. jf

Perspective: 2 frames of 5 points Show graphically that with fewer than 5 points, there is always wiggle room between camera frames

8-point algorithm Longuet-Higgins Hartley normalization

SFM under orthographic projection Orthographic imaging ignores depth: Equation counting: 2FP+F equations (extra eqn. for each frame: set z motion to 0) 3P + 6(F-1) unknowns (same as perspective) But equations are not independent (complicated proof omitted) 2 frames is not enough With 3 frames, need at least 4 points

Orthography: 3 frames of 4 points Show graphically the wiggle room with < 3 frames or < 4 points

Factorization Recall: Stack into measurement matrix: rotation 4xP 2FxP 2Fx4 (Tomasi and Kanade 1992) measurement = motion x shape

Subtracting centroid Place world origin at centroid of points: Then subtract centroid of image coordinates per frame:

Registered measurements This leads to the registered measurement matrix: 3xP 2FxP 2Fx3 registered measurement = rotation x shape

Rank theorem Similarly, Use SVD to enforce rank constraint: This reduces effects of noise in a robust, stable way 3

Euclidean constraints But our choice was arbitrary Solution is unique only up to affine transformation Impose metric constraints to solve for Q: for any invertible 3x3 matrix Q use least squares, then Cholesky decomposition

Algorithm summary Tomasi-Kanade factorization for SFM: (Quadratic equations require nonlinear minimization)

Results

More results

Handling occlusion Unknown image measurement pair (ufp,vfp) in frame f can be reconstructed if p is visible in 3 image frames 3 other points are visible in 4 frames

Occlusion results ping pong ball rotated 450 degrees 84% of data hallucinated from 16%

Factorization extensions Poelman and Kanade (1994): Paraperspective Costeira and Kanade (1995): Multibody factorization Sturm and Triggs (1996): Perspective, fixed rank algorithm to speed computation multibody (Costeira and Kanade) results

Planar parallax See Irani

Using dynamics We have looked at batch methods. Now incremental methods. A. Davison real-time reconstruction

Texture mapping Pollefeys Depth image Triangle mesh Texture image Textured 3D Wireframe model