68
Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry: Homography Epipolar geometry, the essential matrix Camera calibration, the fundamental matrix

Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry: Homography Epipolar geometry, the essential matrix Camera

  • View
    226

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Last Time

Pinhole camera model, projection A taste of projective geometry Two view geometry:

Homography Epipolar geometry, the essential matrix Camera calibration, the fundamental matrix

Page 2: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Epipolar Lines

epipolar linesepipolar linesepipolar linesepipolar lines

BaselineBaselineOO O’O’

epipolar planeepipolar plane

' 0Tp Ep

Page 3: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Stereo Vision

Objective: 3D reconstruction Input: 2 (or more) images taken with calibrated

cameras Output: 3D structure of scene Steps:

Rectification Matching Depth estimation

Page 4: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Rectification

We will assume images have been rectified so that epipolar lines correspond to scan lines Image planes of cameras are parallel. Focal points are at same height. Focal lengths same.

Then, epipolar lines fall along the horizontal scan lines of the images

Any stereo pair can be rectified by rotating and scaling the two image planes (=homography) so that they become parallel to baseline

Page 5: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Rectification

Image Reprojection reproject image planes onto

common plane parallel to baseline Notice, only focal point of camera

really matters(Seitz)

Page 6: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Cyclopean Coordinates

Origin at midpoint between camera centers Axes parallel to those of the two (rectified) cameras

( / 2),

( / 2),

( ) ( ),

2 2

l l

r r

l r

l r l r

l r l r l r

f X b fYx y

Z Zf X b fY

x yZ Z

fbx x

Zb x x b y y fb

X Y Zx x x x x x

Page 7: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Disparity

The difference is called “disparity” d is inversely related to Z: greater sensitivity to

nearby points d is directly related to b: sensitivity to small

baseline

l r

fbZ

x x

l rd x x

Page 8: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Main Step: Correspondence Search What to match?

Objects?

More identifiable, but difficult to compute Pixels?

Easier to handle, but maybe ambiguous Edges? Collections of pixels (regions)?

Page 9: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Matching objects vs. Pixels

Left Right

scanline

Page 10: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Random Dot Stereogram

Using random dot pairs Julesz showed that recognition is not needed for stereo

Page 11: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Random Dot in Motion

Page 12: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Finding Matches

Under what conditions pixels can be matched? Ignoring specularities, we can assume that matching pixels

have the same brightness (constant brightness assumption)

Still, changes in gain and sensitivity may change the values of pixels

Common solution: Use larger windows Normalized correlation

Pros and cons: Small window: accurate match is more likely Large window: fewer candidates

We need a method to eliminate false matches

Page 13: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Window Size

W = 3 W = 20

Page 14: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Constraining the Search

Restrict search to epipolar lines (1D search) Use larger elements (larger windows, edges,

regions)

Problem: large elements may be distorted

Enforce smoothness

Problem: discontinuities at object boundaries

Enforce ordering

Problem: not always true

Page 15: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

1D Search

Page 16: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

SSD error

disparity

1D Search More efficient Fewer false matches

Page 17: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Ordering

Page 18: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Ordering

Page 19: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Correspondence as Optimization Most stereo algorithms attempt to minimize a

functional that usually consists of two terms:

where

- penalizes for quality of a match (unary)

- penalizes non smooth (or even non fronto-parallel) reconstructions (binary)

Many different optimization approaches were proposed

match data smoothnessE E E

smoothnessEdataE

Page 20: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Comparison of Stereo Algorithms

D. Scharstein and R. Szeliski. "A Taxonomy and Evaluation of Dense Two-Frame Stereo Correspondence Algorithms," International Journal of Computer Vision, 47 (2002), pp. 7-42.

Ground truthScene

Page 21: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Scharstein and Szeliski

Page 22: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Results with window correlation

Window-based matching(best window size)

Ground truth

Page 23: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cuts

Ground truthGraph cuts

Page 24: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Stereo Algorithms

We’ll briefly review several algorithms: Dynamic programming Minimal cut/Max flow Space carving Graph cut optimization

Page 25: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

?

Page 26: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

1D Methods: Dynamic Programming Discretize the 3-D space Find the correct curve at every slice

(A slice = epipolar plane)

Page 27: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Dynamic programming

Find correspondences of each epipolar

line separately

Page 28: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Dynamic programming

Page 29: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Dynamic programming

How do we find the best curve? Assign weight of all edges

insertion

matchdeletion

Page 30: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Dynamic programming

How do we find the best curve? Assign weight of all edges Find shortest path

Dijkstra

insertion

matchdeletion

Page 31: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Results

Page 32: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Dynamic programming

Advantages Simple, efficient Globally optimal

Disadvantages Each slice computed independently

(smoothness is not enforced between slices) Problems due to discretization (tilted planes)

Page 33: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Page 34: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Page 35: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Page 36: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Page 37: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Objective: find the optimal cut using all the slices simultaneously.

Page 38: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Construct a graph: Every voxel (3-D point in space) is a node Every node is connected to its 6 neighbors

Page 39: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Weights on the edges: Data cost: change in pixel value

data

data

Page 40: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Weights on the edges: Data cost: change in pixel value Smoothness cost: change in depth

smooth

smooth

smooth

smooth

Page 41: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Weights on the edges: Data cost: change in pixel value Smoothness cost: change in depth

Page 42: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Add source and sink Find min cut

Source

Sink

Page 43: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Data penalty

Smoothness penalty

Page 44: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Results

Input Min cut Dynamic programming

Page 45: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Min Cut/Max Flow

Advantages All slices are optimized simultaneously Efficient

Disadvantages Extension to multi-camera is difficult Discretization

Page 46: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Multi-view stereo Every point in space

corresponds to a match in the images

Compute data term for each match

Space Carving

Page 47: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

0.5 0.4 0.8 0.9 0.9 0.8 0.9 0.3 0.2

Space Carving

Multi-view stereo Every point in space

corresponds to a match in the images

Compute data term for each match (“photo-consistency”)

Page 48: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Space Carving

Dynamic data term (taking occlusion into account)

Order of sweep is important

Page 49: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Space Carving

Page 50: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Space Carving

Done for all slices simultaneously

Page 51: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Space Carving

Done for all slices simultaneously

Page 52: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Space Carving

Done for all slices simultaneously

Page 53: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Space Carving

Computes a bound on the object, the visual hull More camera views: better result

Page 54: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Space Carving: Results

Page 55: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Space Carving: Results

Page 56: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Space Carving

Advantages True multi-views stereo Handles occlusion

Disadvantages Limited to visual hull Lacks smoothness term Noise may introduce holes,

allowing for noise may thicken shape

Discretization

Page 57: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cut Optimization

Stereo is a minimization problem

Possible solution: local search (gradient descent) Problem: inefficient, local minima Instead, search larger areas at every iteration

match data smoothnessE E E

Page 58: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cut Optimization

Construct a graph to represent the problem: Nodes:

Pixels (in first image) k discrete depth values

Edges: From every pixel node to a

depth node (data term) Neighboring nodes (smoothness)

Assign weights corresponding to pixel intensities to get a global cost function

pixels

depths

…21 k

Page 59: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cut Optimization

Objective: Multiway cut Edges:

Every pixel remains connected to one depth node

Edges between neighboring nodes only if they are connected to same depth node

Nodes are assigned the depth that they are connected to

Multiway cut is NP-complete, solve iteratively

……21 k3

pixels

depths

Page 60: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cut Optimization

α-β swap Nodes labeled α or β, (i.e.,

connected to or )

can change their labeling to α or β

Edges between neighbors are updated according to the new labeling

Other edges are not changed Finding best swap = min cut!

α β

……21 k3

pixels

depths

Page 61: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cut Optimization

Example: 1-2 swap

…… k

… k1 2 1 23 3

Page 62: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cut Optimization

Example: 1-2 swap

… k1 2 3 … k321

Connect the nodes labeled 1 or 2 to both

labels

Page 63: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cut Optimization

Example: 1-2 swap… k3

2

1 … k3

2

1

Mark 1 as source and 2 as sink Find minimal cut

Page 64: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cut Optimization

Example: 1-2 swap

… k3

2

1

… k1 2 3

Erase edges that were on the cut

Result: a new labeling of the 1,2 nodes

Page 65: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cut Optimization

Start with an arbitrary labeling For every pair {α, β} є {1,…,k}

Find the best α-β swap (minimizing the function) Update the graph (add and erase edges)

Quit when no pair improves the cost function Induce pixel labels

Page 66: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Graph Cuts: Results

Page 67: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

…21 k3

Advantages State of the art results Efficient Bound on approximation quality Same technique can be applied to other

problems (e.g., image restoration)Disadvantages Discretization Occlusion Still room for improvement

Graph Cut Optimization

Page 68: Last Time Pinhole camera model, projection A taste of projective geometry Two view geometry:  Homography  Epipolar geometry, the essential matrix  Camera

Summary

Stereo vision: shape reconstruction from two or more images

Steps: Rectification Correspondence search Depth estimation

Algorithms: Dynamic programming Min cut/max flow Space carving Graph cuts