Augmented Reality: Beyond the Hype

  • 7,313 views
Uploaded on

Talk given as part of BBytes series at InfoLab21

Talk given as part of BBytes series at InfoLab21

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
No Downloads

Views

Total Views
7,313
On Slideshare
0
From Embeds
0
Number of Embeds
7

Actions

Shares
Downloads
236
Comments
2
Likes
14

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. AUGMENTED REALITY BEYOND THE HYPE Dr Paul Coulton and Klen Čopič PuciharBanksy
  • 2. Mixed Reality Real Augmented Augmented VirtualEnvironment Reality Virtuality RealityMIXED REALITY CONTINUUM Paul Milgram
  • 3. WHAT ISAUGMENTED REALITY?The term AR is being used in all sorts of ways but thegenerally accepted definition is that it:Combines the real and virtual Is interactive in real time Is registered in 3D
  • 4. WHAT ISNT AUGMENTED REALITY? Location‐based services   Barcode detection (QR codes)   Augmenting still images Special effects in movies   Photo‐based object recognitionAlessandro Mulloni ... Yu-Gi-Oh! Zexal
  • 5. GENERAL CHALLENGES OF AR Strict real time operation (30Hz) High spatial precision  (1cm, 1 degree) Robustness for operation by human userAlessandro Mulloni
  • 6. CHALLENGESOF MOBILE ARSame level of performance  as desktop AR  No unrealistic assumptions  about hardwareVariable operating contexts Layar
  • 7. Image Denno CoilFLAVOURS OF MOBILE AR WEARABLE AR Wearable system Head‐mounted display AR always in view  (immersive) Demo Sixth Sense
  • 8. FLAVOURS OF MOBILE AR HANDHELD AR Mobile phone as platformPhone acts as a “magic lens”   Non‐immersive view Tom’s Hardware
  • 9. WHY USE PHONES?Low cost, Ubiquity, Robust, Self Contained
  • 10. WHY NOT USE PHONES?Low memory, Limited power, Small Screen, Limited Inputs available, Fragmentation
  • 11. PERCEPTUAL PROBLEMS You see through the camera not the phone!
  • 12. Get video frame from camera DEVELOPING AR Estimate position and orientation of the cameraAPPLICATIONS Render the augmented scene (video and virtual) What is involved in the process? Render GUI Process the user input Update application status
  • 13. Phones withHandheld Phones acceleromet AR with GPS ers Phones withDisplays two Camera cameras Phones PDA Phone Phones On device with 3D with Wearable HW compass AR AR Phone Thin Phones with PDA client AR gyroscopes Thin client AR1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 EVOLUTION OF MOBILE AR
  • 14. ESTIMATINGTHE DEVICE POSE Sensor trackingVision‐based tracking
  • 15. SENSOR BASED Used by many commercial “AR browsers” GPS, Compass,Accelerometer, (Gyroscope) Demo Wikitude
  • 16. GPSOriginally 24 satellites but 31 as of 2008Bill Clinton cleared GPS for commercial use in 1996
  • 17. GPS Spatial ScatteringSystem and Environmental Effects
  • 18. ACCELEROMETERS
  • 19. ACCELEROMETERS
  • 20. MAGNETOMETERS
  • 21. David McCandlessInformation is Beautiful DATA SOURCES Quality, Availability, Crowd Sourcing
  • 22. VISION BASED MARKER TRACKING Standard Vision techniques Marker provides 4 cornersfrom this we can get the pose
  • 23. VISION BASED MARKER TRACKING1.Convert image to black and white 2. Search for edges3. Follow edges 4. Find rectangle corners !* K* C* :* Daniel Wagner
  • 24. VISION BASED MARKER TRACKING5.Estimate homography using 4corners 6. Extract pattern by sampling7. Check Pattern Daniel Wagner
  • 25. MARKER TRACKING PIPELINE Daniel Wagner
  • 26. VISION BASED MARKER TRACKING Nintendo 3DS AR
  • 27. VISION BASED NATURAL FEATURE TRACKINGTracking features from natural environment More difficult than marker Less established techniques Slower than marker based PTAM
  • 28. VISION BASED NATURAL FEATURE TRACKING Edges - boundariesCorners - local 2D structure Blobs - regions rather than pointsRidges - elongated structures
  • 29. VISION BASED NATURAL FEATURE TRACKING OfflineUse still images to build data base of features
  • 30. Camera ImageVISION BASED Keypoint Detection NATURAL FEATURE Outlier Removal TRACKING Pose Estimation OnlineCreating maps are markers RECOGNITION online Real-Time Pipeline Pose
  • 31. VISION BASED NATURAL FEATURE TRACKING Online
  • 32. • SENSORS • VISION • Noise Output • High Accuracy • Low Accuracy • Local Pose • Global Pose • Memory Intensive • Works when • Works only when nothing to track Objects to Track HYBRID TECHNIQUES
  • 33. DESIGN MOCKUPSAR is easy to understand, but hard to explain in words. Ease of creation. Holding strong memorable message.
  • 34. CONCLUSIONS• Augmented Reality is currently the hot topic from a continuum of possible systems.• Mobile phones provide the most obvious platform for the widespread adoption of AR.• Majority of current commercial offering are sensor based which gives crude contextual sensitivity but this may be fine for your application. These solutions generally are dependant on quality of the data source.• Marker based solutions give greater accuracy although we arent likely to cover the world with markers work well for advertising.• Marker-less systems offer the ‘dream’ of AR but currently present considerable technical challenges.• If your considering AR ask what benefits it gives the user apart from being ‘cool’.
  • 35. OTHER INTERESTING AR STUFF