1 / 42

Ilya Landa 03/30/2008

Summary of “Data Processing Algorithm for Generating 3D building Facade Meshes From Laser Scans and Camera Images”. An article by Christian Frueh, Siddharth Jain, Avideh Zakhor. Ilya Landa 03/30/2008. Overview. The need for creating 3D models of urban environments Current methods

Download Presentation

Ilya Landa 03/30/2008

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Summary of“Data Processing Algorithm for Generating 3D building Facade Meshes From Laser Scans and Camera Images”. An article by Christian Frueh, Siddharth Jain, Avideh Zakhor Ilya Landa 03/30/2008

  2. Overview • The need for creating 3D models of urban environments • Current methods • Aerial Photography • Vision Based Methods • Robot Mounted 2D/3D Lasers • Van Mounted Lasers Equipped With GPS

  3. Experiment Setup • Vertical 2D laser • Horizontal 2D laser • Digital Camera

  4. Experiment Setup

  5. Problems and Complications

  6. Raw Triangulation

  7. Occlusion

  8. Reverse Order Scans

  9. Reflective or Transparent Surfaces

  10. Organizing Scan Points

  11. Downtown Assumptions • Buildings have relatively flat facades that are parallel to the road, thus perpendicular to the scans • Ground is a horizontal plane in all scans • Surveyed landscapes are not dominated by trees

  12. Depth Levels • Main Depth – most frequent value • Split Depth – local minimum • Background Layer – around main depth • Foreground Layer – trees, cars, etc. • Ground Level 

  13. Accumulative Histogram

  14. Results of the Split Raw Points Foreground Layer Horizontal Laser Points Added Background Layer + Ground Points

  15. Projecting Foreground Objects onto the background and ground layers

  16. Removing Window Holes • Scanning a window results in a series of points with a random and large depth. • If such points are in between points on the main depth, the hole can be eliminated.

  17. Point Improvement Overview

  18. Texture Generation

  19. Camera Data

  20. Naive Foreground Removal

  21. Background / Foreground SplittingSetup

  22. Background / Foreground SplittingNo Obstacles

  23. No Obstacles Example

  24. Background / Foreground SplittingSmall Obstacle

  25. Small Obstacle Example

  26. Background / Foreground SplittingLarge Obstacle

  27. Large Obstacle Example

  28. Split Results

  29. Combining Split Photos

  30. Linear Hole Filling

  31. Copy-Paste

  32. Final Texture Atlas

  33. Quality Comparison

  34. Raw vs. ImprovedVisual Comparison of 73 City Block Fronts

  35. “Significantly Better” Examples

  36. “Better” Examples

  37. “Bad” Example

  38. Downtown Berkeley

  39. Downtown Berkeley

  40. Downtown Berkeley

  41. Downtown Berkeley

More Related