Text of 3D photography Marc Pollefeys Fall 2008 http://www.inf.ethz.ch/personal/pomarc/courses/3dphoto
Slide 1
3D photography Marc Pollefeys Fall 2008 http://www.inf.ethz.ch/personal/pomarc/courses/3dphoto/
Slide 2
3D Photography Obtaining 3D shape (and appearance) of real-world objects
Slide 3
Motivation Applications in many different areas A few examples
Slide 4
Surgery - simulation simulate results of surgery allows preoperative planning
Slide 5
Surgery - teaching Capture models of surgery for interactive learning
Slide 6
Clothing Scan a person, custom-fit clothing
Slide 7
Forensics Crime scene recording and analysis
Slide 8
Forensics
Slide 9
3D urban modeling UNC/UKY UrbanScape project
Slide 10
Industrial inspection Verify specifications Compare measured model with CAD
Slide 11
Scanning industrial sites as-build 3D model of off-shore oil platform
Slide 12
Robot navigation small tethered rover pan/tilt stereo head ESA project our task: Calibration + Terrain modelling + Visualization
Slide 13
Architecture Survey Stability analysis Plan renovations
Slide 14
Architecture Survey Stability analysis Plan renovations
Slide 15
Cultural heritage Virtual Monticello Allow virtual visits
Slide 16
Cultural heritage Stanfords Digital Michelangelo Digital archive Art historic studies
Slide 17
Archaeology accuracy ~1/500 from DV video (i.e. 140kb jpegs 576x720)
Slide 18
Record different excavation layers Archaeology Generate & verify construction hypothesis Layer 1 Layer 2 Generate 4D excavation record
Slide 19
Computer games
Slide 20
Course objectives To understand the concepts that allow to recover 3D shape from images Explore the state of the art in 3D photography Implement a 3D photography system/algorithm
Slide 21
Material Slides and more http://www.inf.ethz.ch/personal/pomarc/courses/3dphoto/ Also check out on-line shape-from-video tutorial: http://www.cs.unc.edu/~marc/tutorial.pdf http://www.cs.unc.edu/~marc/tutorial/ Other interesting stuff: Book by Hartley & Zisserman, Multiple View Geometry Peter Allens 3D photography class webpage http://www1.cs.columbia.edu/~allen/F07/schedule.html http://www1.cs.columbia.edu/~allen/F07/schedule.html
Slide 22
Learning approach Lectures: cover main 3D photography concepts and approaches. Exercice sessions: Assignments (1 st half semester) Paper presentations and project updates (2 nd half semester) 3D photography project: Choose topic, define scope (1 st half semster) Implement algorithm/system (2 nd half semester) Grade distribution Assignments & paper presentation: 50% Course project: 50%
Slide 23
Content camera model and calibration single-view metrology triangulation epipolar geometry, stereo and rectification structured-light, active techniques feature tracking and matching structure-from-motion shape-from-silhouettes space-carving
Slide 24
3D photography course schedule (tentative) LectureExercise Sept 17Introduction- Sept 24Geometry & Camera modelCamera calibration Oct. 1Single View MetrologyMeasuring in images Oct. 8Feature Tracking/matching (Jan-Michael Frahm?) Correspondence computation Oct. 15Epipolar Geometry (David Gallup?) F-matrix computation Oct. 22Shape-from-SilhouettesVisual-hull computation Oct. 29Stereo matchingProject proposals Nov. 5Structured light and active range sensing (TBD) Papers Nov. 12Structure from motionPapers Nov. 19Multi-view geometry and self-calibration Papers Nov. 26Shape-from-XPapers Dec. 33D modeling and registrationPapers Dec. 10Appearance modeling and image-based rendering (TBD) Papers Dec. 17Final project presentations
Slide 25
Fast Forward! Quick overview of what is coming
Slide 26
Camera models and geometry Pinhole camera Geometric transformations in 2D and 3D or
Slide 27
Camera Calibration Know 2D/3D correspondences, compute projection matrix also radial distortion (non-linear)
Slide 28
Single View Metrology
Slide 29
Antonio Criminisi
Slide 30
Feature tracking and matching Harris corners, KLT features, SIFT features key concepts: invariance of extraction, descriptors to viewpoint, exposure and illumination changes
Slide 31
l2l2 3D from images C1C1 m1m1 M? L1L1 m2m2 L2L2 M C2C2 Triangulation - calibration - correspondences
Slide 32
Epipolar Geometry Fundamental matrix Essential matrix Also how to robustly compute from images
Slide 33
Stereo and rectification Warp images to simplify epipolar geometry Compute correspondences for all pixels
Slide 34
Structured-light Projector = camera Use specific patterns to obtain correspondences
Slide 35
Initialize Motion (P 1,P 2 compatibel with F) Initialize Structure (minimize reprojection error) Extend motion (compute pose through matches seen in 2 or more previous views) Extend structure (Initialize new structure, refine existing structure) Structure from motion
