[06-30213][06-30241][06-25024]
Computer Vision and Imaging &
Robot Vision
Dr Hyung Jin Chang Dr Yixing Gao
h.j.chang@bham.ac.uk y.gao.8@bham.ac.uk
School of Computer Science
MULTIPLE-VIEW GEOMETRY
(CHAPTERS 7, 8 (FORSYTH & PONCE) CHAPTER 11 (R. SZELISKI))
Single
• Isitpossibletoinferatrue3Dscenestructurefrom a single image?
–
view geometry
Well, not really
• Structureanddepthareambiguous…
Slide credit: Svetlana Lazebnik, Kristen Grauman
The reason behind depth ambiguity
• Allpointsalongaraythatpassesthroughacamera center are projected into the same point in the image plane.
• Impossibletoinfer3Dpointfromasingle2Dpoint (without prior on the scene structure, that is)
𝑿?
𝑿? 𝒙
𝑿?
Taking advantage of ambiguity
• Anamorphosis(earlierthan15thcentury)
Edgar Mueller
Julian Beever
Take two images = Stereo!
• Mucheasierusingapairofviews…
http://www.well.com/~jimg/stereo/stereo_list.html
Slide credit: Kristen Grauman
What is stereo?
• Genericformulation:givenseveralimagesofthe same object or scene, compute a representation of its 3D shape
Slide credit: Svetlana Lazebnik, Steve Seitz
What is stereo?
• Genericformulation:givenseveralimagesofthe same object or scene, compute a representation of its 3D shape
• “Images of the same object or scene”
• Arbitrary number of images (from two to thousands)
• Arbitrary camera positions (isolated cameras or video sequence)
• “Representation of 3D shape” • Depth maps
• Point clouds
• Patch clouds
• Meshes
• Volumetric models
Slide credit: Svetlana Lazebnik, Steve Seitz
What is stereo?
• Genericformulation:givenseveralimagesofthe same object or scene, compute a representation of its 3D shape
How do we
combine
multiple views?
Require a cross-view calibration!
Slide credit: Svetlana Lazebnik, Steve Seitz
What is stereo?
• Genericformulation:givenseveralimagesofthe same object or scene, compute a representation of its 3D shape
How do we
combine
multiple views?
Require a cross-view calibration!
Slide credit: Svetlana Lazebnik, Steve Seitz
What is stereo?
• Genericformulation:givenseveralimagesofthe same object or scene, compute a representation of its 3D shape
• BuildingRomeinaday http://grail.cs.washington.edu/rome/
The data set consists of 150,000 images from Flickr.com associated with the tags “Rome” or “Roma”. Matching and reconstruction took a total of 21 hours on a cluster with 496 compute cores. Upon matching, the images organized themselves into a number of groups corresponding to the major landmarks in the city of Rome.
Slide credit: Svetlana Lazebnik, Steve Seitz
What is stereo?
• Narrowerformulation:givenacalibratedbinocular stereo pair, fuse it to produce a depth image
Image1
Image 2
Dense depth map
Slide credit: Svetlana Lazebnik, Steve Seitz
What is stereo?
• Narrowerformulation:givenacalibratedbinocular stereo pair, fuse it to produce a depth image
• Peoplecantothis…
Stereograms: Sir Charles Wheatstone, 1838
Slide credit: Svetlana Lazebnik, Steve Seitz
What is stereo?
• Narrowerformulation:givenacalibratedbinocular stereo pair, fuse it to produce a depth image
• Peoplecantothis…
Anaglyph
Polarized Glasses
Shutter
Slide credit: Svetlana Lazebnik, Steve Seitz
What is stereo?
• Narrowerformulation:givenacalibratedbinocular stereo pair, fuse it to produce a depth image
• Peoplecantothis…
Autostereograms: www.magiceye.com
Slide credit: Svetlana Lazebnik, Steve Seitz
What is stereo?
• Narrowerformulation:givenacalibratedbinocular stereo pair, fuse it to produce a depth image
• Peoplecantothis…
Autostereograms: www.magiceye.com
Slide credit: Svetlana Lazebnik, Steve Seitz
Application: autonomous robots
Nomad robot looks for meteorites Real-time stereo system on Mars on Antarctica
Application: autonomous robots
Mars Exploration Rovers, NASA, 2003
Courtesy NASA/JPL-Caltech
Applications: Stereo Cameras
Thank you
Hyung Jin Chang
15/03/2021