Course 7 Motion.

Slides:



Advertisements
Similar presentations
More Vectors.
Advertisements

Announcements. Structure-from-Motion Determining the 3-D structure of the world, and/or the motion of a camera using a sequence of images taken by a moving.
Computer vision: models, learning and inference
Mapping: Scaling Rotation Translation Warp
Epipolar geometry. (i)Correspondence geometry: Given an image point x in the first view, how does this constrain the position of the corresponding point.
Uncalibrated Geometry & Stratification Sastry and Yang
Ch. 4: Velocity Kinematics
3-D Geometry.
Introduction to Robotics Lecture II Alfred Bruckstein Yaniv Altshuler.
Solution Sets of Linear Systems (9/21/05)
© 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.
The Pinhole Camera Model
CS223b, Jana Kosecka Rigid Body Motion and Image Formation.
3D Motion Estimation. 3D model construction Video Manipulation.
Camera parameters Extrinisic parameters define location and orientation of camera reference frame with respect to world frame Intrinsic parameters define.
Linear Algebra Review CSE169: Computer Animation Instructor: Steve Rotenberg UCSD, Winter 2005.
Linear Algebra – Linear Equations
Length and Dot Product in R n Notes: is called a unit vector. Notes: The length of a vector is also called its norm. Chapter 5 Inner Product Spaces.
Linear Systems of Equations
ECON 1150 Matrix Operations Special Matrices
Euclidean cameras and strong (Euclidean) calibration Intrinsic and extrinsic parameters Linear least-squares methods Linear calibration Degenerate point.
1 Preview At least two views are required to access the depth of a scene point and in turn to reconstruct scene structure Multiple views can be obtained.
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.
Course 12 Calibration. 1.Introduction In theoretic discussions, we have assumed: Camera is located at the origin of coordinate system of scene.
Geometric Models & Camera Calibration
Rotations and Translations
Copyright © 2013, 2009, 2005 Pearson Education, Inc. 1 5 Systems and Matrices Copyright © 2013, 2009, 2005 Pearson Education, Inc.
POSITION & ORIENTATION ANALYSIS. This lecture continues the discussion on a body that cannot be treated as a single particle but as a combination of a.
视觉的三维运动理解 刘允才 上海交通大学 2002 年 11 月 16 日 Understanding 3D Motion from Images Yuncai Liu Shanghai Jiao Tong University November 16, 2002.
Geometric Camera Models
Systems of Equations and Inequalities Systems of Linear Equations: Substitution and Elimination Matrices Determinants Systems of Non-linear Equations Systems.
Elementary Linear Algebra Anton & Rorres, 9th Edition
Chap. 5 Inner Product Spaces 5.1 Length and Dot Product in R n 5.2 Inner Product Spaces 5.3 Orthonormal Bases: Gram-Schmidt Process 5.4 Mathematical Models.
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)
2.1 – Linear and Quadratic Equations Linear Equations.
This Week WeekTopic Week 1 Coordinate Systems, Basic Functions Week 2 Trigonometry and Vectors (Part 1) Week 3 Vectors (Part 2) Week 4 Vectors (Part 3:
Camera Model Calibration
Determining 3D Structure and Motion of Man-made Objects from Corners.
Digital Image Processing Additional Material : Imaging Geometry 11 September 2006 Digital Image Processing Additional Material : Imaging Geometry 11 September.
Inner Product Spaces Euclidean n-space: Euclidean n-space: vector lengthdot productEuclidean n-space R n was defined to be the set of all ordered.
CS682, Jana Kosecka Rigid Body Motion and Image Formation Jana Kosecka
7.3 Linear Systems of Equations. Gauss Elimination
Review of Linear Algebra
Lecture 3 Jitendra Malik
Chapter 2 Planes and Lines
1 Equations, Inequalities, and Mathematical Modeling
Computer Graphics 3D Transformations
Solving Systems of Equations in Three Variables
Homogeneous Coordinates (Projective Space)
3D Motion Estimation.
Epipolar geometry.
Linear Inequalities and Absolute Value
Structure from motion Input: Output: (Tomasi and Kanade)
A quadratic equation is written in the Standard Form,
Numerical Analysis Lecture 16.
Example 1: Finding Solutions of Equations with Two Variables
Reconstruction.
2D Geometric Transformations
Equations of Straight Lines
Linear Algebra Lecture 39.
3.5 Solving Nonlinear Systems
Single-view geometry Odilon Redon, Cyclops, 1914.
CSCE441: Computer Graphics 2D/3D Transformations
Section Solving Linear Systems Algebraically
Example 2B: Solving Linear Systems by Elimination
Matrices are identified by their size.
The Pinhole Camera Model
Eigenvalues and Eigenvectors
Structure from motion Input: Output: (Tomasi and Kanade)
Presentation transcript:

Course 7 Motion

Course 7 Motion 3D motion expression: Motion of a rigid object in 3D is usually expressed as a rotation around system origin followed by a translation. Let be a 3D point of an object at time Let be the point after motion at

be rotation , a 3x3 orthonormal matrix. be translation vector. Then, from time t1 to t2 :

r112+r122+r132=1 r212+r222+r232=1 r312+r322+r332=1 r11r21+r12r22+r13r23=0 r21r31+r22r32+r21r33=0 r31r11+r32r12+r33r13=0

Note: There are only 3 independent parameters in R. Properties of Rotation matrix: (1) (2) (3) (4)

2) R is expressed by rotation axis and rotation angle. Let rotation axis be where: n12+n22+n32=1. Rotation is made by a rotating with an angle  around a rotation axis. 

Then the elements of R are:

3) Express R by 3 rotation angles.

4) Quaternion form of rotation: Quaternion is a four-element vector, which can be used to express a rotation: Let a rotation around axis by angle 

Then: Where -----scalar part -----vector part Quaternion conjugation

Quaternion product: A 3D vector can be expressed as a quaternion with scalar part being zero: Pure rotation in 3D: Express the rotation by quaternion:

5) Homogenous coordinate systems: Then in 3D coordinate system is written as in homogenous coordinate system.

2. Motion from 3D PCs: (1) method 1: 3 or more point on a rigid object (at least 3 point not collinear) can be seen over two-time instances t1 and t2: (1) (2) (3) Subtract (1) from (2), (3) respectively: (4) (5)

(5): (6) From (4),(5) and (6), R can be solved (never forget rotation matrix constraints). After finding R, (2) method 2: From eq (4), vector to subjects to a pure rotation

The vector is perpendicular to rotation axis The same is true for

Thus, And rotation angle can be determined by the angle of two planes At least 3 not collinear 3D points on rigid object is needed to determine 3D motion!

3. Motion from 2D PC: ------- From 2D images to determine 3D motion. ------- At least 5-point correspondences over two-image view are required. ------- 3D translation can only be determined over a scale factor. ------- Degeneration case: 3D points are on a quadratic surface.

Assume: ------ Single stationary camera. ------ Central projection model. ------ Rigid moving object. ------ Focus length f = 1, thus 3D point: 2D image:

Let 3D motion from of time to of (1) Where From equ (1) (2) Apply to both sides of equ (2) (3)

Apply to both sides of equ (3) (4) Let we define (5) The eq(4) can be rewritten as (6) Note: eq.(6) is a homogeneous scalar equation. is a matrix containing only motion parameters, 8 or more PCs can uniquely determine E, subject to:

After matrix E is found ,translation can be solved: Or (7) can be determined from eq(7) subject to

Once is obtained, rotation R can be obtained by least-square method: (8) Or let

Note, 180o reflection of motion is still a solution of equ (7) (homogeneous equation). In this case, object is moving behind the camera. To check for a real solution, we apply to both sides of equ (2). Therefore if z > 0, it must hold that Thus if, let  

Remark: at least 8 non-degenerated PC’s over two image frames are needed to solve for a 3D motion using a linear method! 4. Motion from LC’s: ------ from two image frames of a single camera, 3D motion can never be solved. ------ over 3 frames, at least 6 LC’s are required. ------ motion models

Model A

Model B Relation between model A and B:

Now, let we consider model the case B: At time t1: (10) At time t2 : (11)

At time t3: (12) From equ. (11) (13) Applying to both sides of equ. (13) and notice that

We get (14) In the same way (15) Eliminate from eqs (14) and (15), we obtain: (16) If we define (17)

F, G, H are 3x3 matrices. Then equ. (16) can be written in compact form: (18) Where Note: equ. (18) is a vector equation containing 3 linear homogeneous equations. And only two of them are linear independent (prove it).

Therefore, 13 LC’s over 3 frames are needed to linearly solve for F, G, and H.  Let we define: We have After E is found, translation can be solved by: Subject to , let be

Similarly, we define Then And Subject to , let the solution be

In solving for R12 and R13, we rather reconstruct E and E’ for consistence (remember E and E’ were column by column). are chosen such that Rotations can then be solved by:

Remark: check revered rotations: if Next, we determine relative amplitudes of translations Let substitute them into eqs (17):

when m, n are solved, translations are: Build structures of 3D lines: Direction of 3D line : Position of : Choose sign to make

Check translation reflection. Evidently, 3D Line Check translation reflection. Evidently, So for and For and  

5. Motion from other image clues: If Else if 5. Motion from other image clues: ------ Optical flow ------ Texture ------ Other clues Will be discussed later.