Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested,...

18
Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see http://www.unc.edu/courses/2004fall/comp/290/089/papers/Triggs- ijcv95.pdf )
  • date post

    21-Dec-2015
  • Category

    Documents

  • view

    215
  • download

    0

Transcript of Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested,...

Page 1: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Triangulation and Multi-View Geometry

Class 9

Read notes Section 3.3, 4.3-4.4, 5.1(if interested, read Triggs’s paper on MVG using tensor notation, see

http://www.unc.edu/courses/2004fall/comp/290/089/papers/Triggs-ijcv95.pdf

)

Page 2: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Step 1. Extract featuresStep 2. Compute a set of potential matchesStep 3. do

Step 3.1 select minimal sample (i.e. 7 matches)

Step 3.2 compute solution(s) for F

Step 3.3 determine inliers

until (#inliers,#samples)<95%

samples#7)1(1

matches#inliers#

#inliers 90%

80%

70% 60%

50%

#samples

5 13 35 106 382

Step 4. Compute F based on all inliersStep 5. Look for additional matchesStep 6. Refine F based on all correct matches

(generate hypothesis)

(verify hypothesis)

Automatic computation of F

Page 3: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Abort verification early

Given n samples and an expected proportion of inliers p, how likely is it that I have observed less than T inliers?abort if P<0.02 (initial sample most probably contained outliers)

(inspired from Chum and Matas BMVC2002)

OOOOOIOOIOOOOOIOOOOOOOIOOOOOIOIOOOOOOOOOIOIIIIOIIIOIOIIIIOOIOIIIIOIOIOIIIIIIII

(use normal approximation to binomial)

To avoid problems this requires to also verify at random!

(but we already have a random sampler anyway)

Page 4: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

restrict search range to neighborhood of epipolar line (e.g. 1.5 pixels)

relax disparity restriction (along epipolar line)

Finding more matches

Page 5: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

• Degenerate cases• Planar scene• Pure rotation

• No unique solution• Remaining DOF filled by noise• Use simpler model (e.g. homography)

• Solution 1: Model selection (Torr et al., ICCV´98, Kanatani, Akaike)

• Compare H and F according to expected residual error (compensate for model complexity)

• Solution 2: RANSAC• Compare H and F according to inlier count

(see next slide)

Degenerate cases:

Page 6: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

RANSAC for (quasi-)degenerate cases

• Full model (8pts, 1D solution)

Sample for out of plane points among outliers

closest rank-6 of Anx9 for all plane inliers

(accept inliers to solution F)

(accept inliers to solution F1,F2&F3)

• Planar model (6pts, 3D solution)

Accept if large number of remaining inliers• Plane+parallax model (plane+2pts)

80% in plane 2% out plane18% outlier

Page 7: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

• Absence of sufficient features (no texture)• Repeated structure ambiguity

(Schaffalitzky and Zisserman, BMVC‘98)

• Robust matcher also finds Robust matcher also finds support for wrong hypothesissupport for wrong hypothesis• solution: detect repetition solution: detect repetition

More problems:

Page 8: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

RANSAC for ambiguous matching

• Include multiple candidate matches in set of potential matches

• Select according to matching probability (~ matching score)

• Helps for repeated structures or scenes with similar features as it avoids an early commitment, but also useful in general

(Tordoff and Murray ECCV02)

Page 9: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

geometric relations between two views is fully

described by recovered 3x3 matrix F

two-view geometry

Page 10: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Triangulation (finally!)

C1x1

L1

x2

L2

X

C2

Triangulation

- calibration

- correspondences

Page 11: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Triangulation• Backprojection

• Triangulation

Iterative least-squares

• Maximum Likelihood Triangulation

C1 x1L1

x2

L2

X

C2

Page 12: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Optimal 3D point in epipolar plane

• Given an epipolar plane, find best 3D point for (m1,m2)

m1

m2

l1 l2l1m1

m2l2

m1´

m2´

Select closest points (m1´,m2´) on epipolar lines

Obtain 3D point through exact triangulationGuarantees minimal reprojection error (given this epipolar plane)

Page 13: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Non-iterative optimal solution

• Reconstruct matches in projective frame by minimizing the reprojection error

• Non-iterative methodDetermine the epipolar plane for reconstruction

Reconstruct optimal point from selected epipolar plane Note: only works for two views

2222

11 ,, MPmMPm DD

(Hartley and Sturm, CVIU´97)

2222

11 ,, lmlm DD (polynomial of degree 6)

m1

m2

l1 l2

3DOF

1DOF

Page 14: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Backprojection

• Represent point as intersection of row and column

Useful presentation for deriving and understanding multiple view geometry(notice 3D planes are linear in 2D point coordinates)

• Condition for solution?

Page 15: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Multi-view geometry

(intersection constraint)

(multi-linearity of determinants)

(= epipolar constraint!)

(counting argument: 11x2-15=7)

Page 16: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Multi-view geometry

(multi-linearity of determinants)

(= trifocal constraint!)

(3x3x3=27 coefficients)

(counting argument: 11x3-15=18)

Page 17: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Multi-view geometry

(multi-linearity of determinants)

(= quadrifocal constraint!)

(3x3x3x3=81 coefficients)

(counting argument: 11x4-15=29)

Page 18: Triangulation and Multi-View Geometry Class 9 Read notes Section 3.3, 4.3-4.4, 5.1 (if interested, read Triggs’s paper on MVG using tensor notation, see.

Next class: rectification and stereo

image I(x,y) image I´(x´,y´)Disparity map D(x,y)

(x´,y´)=(x+D(x,y),y)