3D/4D acquisition. 3D acquisition taxonomy Computer Vision. Computer Vision. 3D acquisition methods. passive. active.

(1)

Computer Vision

3D/4D

acquisition

3D/4D

acquisition

 Computer Vision

Das Bild kann zurzeit nicht angezeigt werden.

3D acquisition taxonomy

uni-directional multi-directional passive uni-directional multi-directional active 3D acquisition methods

Shape-from Stereo Time-of-flight - texture - contour - silhouettes - defocus - shading Line scanning Structured light Photom. stereo

(2)

Computer Vision

3D acquisition taxonomy

Shape-from Stereo Time-of-flight - texture - contour - silhouettes - defocus - shading Line scanning Structured light Photom. stereo Computer Vision

Stereo

(3)

Computer Vision 

(Passive) stereo

: I. A simple configuration Computer Vision 

(Passive) stereo

Simple configuration :

(4)

Computer

Vision

Digital surveying

Computer Vision

A simple stereo setup

(5)

Computer Vision



(Passive) stereo

:

II. General configuration

Computer

Vision

Stereo, the general setup

we start by the relation between the two projections of a point

in the second image the point must be along the projection of the viewing ray for the first camera :

(6)

Computer

Vision

Stereo, the general setup

note that the epipole lies on all the epipolar lines



Computer Vision

Andrea Fusiello, CVonline

Stereo, the general setup

(7)

Computer Vision

Stereo, the general setup

Computer

Vision

Correspondence problem : constraints

Reducing the search space :

 1. Points on the epipolar line

 2. Min. and max. depth ⇒ line segment  3. Preservation of order

 4. Smoothness of the disparity field

(8)

Computer

Vision

Correspondence problem : methods

1. correlation

deformations…

small window ⇒ noise!

large window ⇒ bad localisation 2. feature-based

mainly edges and corners

sparse depth image

3. regularisation methods  Computer Vision

Uncalibrated reconstruction

If the camera translates... From 2 views...

(9)

Computer Vision



Uncalibrated reconstruction

A metric reconstruction is possible

From 3 general views

taken with the same camera parameters...

Computer Vision

Temple of the Masks, Edzna, Mexico

(10)

Computer Vision

Uncalibrated reconstruction

Computer Vision

Uncalibrated reconstruction

In the desert… Dunhuang caves along the Silk Road in China Unesco World Heritage

(11)

Computer Vision

Uncalibrated reconstruction

Computer Vision

Uncalibrated reconstruction

(12)

Computer Vision

www.arc3d.be

Webservice,

free for non-commercial use

Shape-from-stills

Computer Vision

3D acquisition taxonomy

(13)

Computer

Vision

Active triangulation

Computer Vision Triangulation  3D measurements Camera Projector

Active triangulation

(14)

Computer

Vision

Active triangulation

Camera image

Computer

(15)

Computer Vision

Active triangulation

Example 1 Cyberware laser scanners



Desktop model for small objects Medium-sized objects

Body scanner Head scanner

Computer

Vision

Active triangulation

Example 2 Minolta



(16)

Computer Vision

3D acquisition taxonomy

Shape-from Stereo Time-of-flight - texture - contour - silhouettes - defocus - shading Line scanning Structured light Photom. stereo Computer

Vision

_{Structured light}

patterns of a special shape are projected onto the scene

deformations of the patterns yield information on the shape

Focus is on combining a good resolution with a minimum number of pattern projections

(17)

Computer Vision  1 0 0 1 10 0 0 0 1

Serial binary patterns

A sequence of patterns with increasingly fine subdivisions

Yields identifiable lines for only patterns

2

n n

Computer

Vision

Reducing the nmb of projections: colour

Binary patterns

Using colours, e.g. 3,



Interference from object colours... Yields identifiable lines for only patterns

2

n n

(18)

Computer Vision

35

One-shot structured light

Computer Vision

One-shot 3D acquisition

(19)

Computer Vision

Shape + texture often needed

Higher resolution

Texture is also extracted

Computer

Vision

Active triangulation

Recent, commercial example



Kinect 3D camera, affordable and compact solution by Microsoft. Projects a 2D point pattern in the NIR, to make it invisible to the human eye

(20)

Computer Vision

Kinect as one-shot, low-cost scanner

Excerpt from the dense NIR dot pattern:

http://research.microsoft.com/apps/video/default.aspx?i 15

Computer Vision

Structured light

(21)

Computer

Vision James Bond

Die another day

Lara Croft Tomb

Raider

Computer

(22)

Computer

Vision

Face animation – replay + effects

Computer Vision

(23)

Computer

Vision

Phase shift

Computer Vision

1. detect phase from 3 subsequently projected

cosine patterns, shifted over 120 degrees

2. unwrap the phases / additional stereo

3. texture is obtained by summing the 3

images / color camera w. slower integration

)

(

cos

)

(

cos

)

(

cos























R

A

I

R

A

I

R

A

I

b g r

Phase shift

(24)

Computer Vision

































b r g b r b g r

I

A

2

2 tan

arctan

3 



Phase shift

Computer Vision

4D acquisition

(25)

Computer Vision

3D acquisition taxonomy

Shape-from Stereo Time-of-flight - texture - contour - silhouettes - defocus - shading Line scanning Structured light Photom. stereo Computer Vision

Time-of-flight

measurement of the time a modulated light signal needs to travel before returning to the sensor this time is proportional to the distance

waves :

1.radar low freq. electromagnetic 2. sonar acoustic waves 3. optical radar optical waves

working principles : 1. pulsed 2. phase shifts

(26)

Computer Vision

Time-of-flight

Example 1: Cyrax  Example 2: Riegl Computer Vision

Pulsed laser (time-of-flight)

Cyrax

Distance = C x

T

2mm - 6mm

accuracy

No reflectors

needed

÷2

(27)

Computer Vision

Step 1:

Target the structure

Computer

Vision

Step 2:

(28)

Computer

Vision

Step 3:

Model fitting in-the-field

Computer

(29)

Computer

Vision

Comparison Lidar - passive

Image courtesy of Tippett Studio

Computer Vision

3D acquisition taxonomy

(30)

Computer

Vision

Shape-from-silhouettes

Computer

(31)

Computer Vision

Outdoor visual hulls

Computer Vision

3D acquisition taxonomy

(32)

Computer Vision

Photometric stereo

constraint propagation eliminated by using light from different directions

simultaneously when the light sources are given different colours



Computer Vision

(33)

Computer

Vision

Mini-dome for photometric stereo

Computer Vision

(34)

Computer Vision

Combining shape and reflectance On-line scanning

In-hand scanning Opportunistic scanning

Multi-modal scanning (multi-spectral, haptics, …) Semantic 3D

Off-the-shelf components / consumer products

Computer Vision

(35)

Computer Vision

There also is surface reflectance (mini-dome).

Computer Vision

(36)

Computer Vision

Procedural modeling

Computer Vision

(37)

Computer Vision

Inverse procedural modeling

Computer Vision

(38)

Computer Vision

Strongest 3D cues for us are 2D...

Computer Vision