Visual Analysis and Semantic Exploration of Urban LIDAR Change Detection - PowerPoint PPT Presentation

About This Presentation
Title:

Visual Analysis and Semantic Exploration of Urban LIDAR Change Detection

Description:

Journal of Photogrammetry and Remote ... Intuitive presentation of changes Non-realistic LOD for scale-appropriate data abstraction Analytical and ... – PowerPoint PPT presentation

Number of Views:198
Avg rating:3.0/5.0
Slides: 29
Provided by: tjbu6
Learn more at: http://www.cs.tufts.edu
Category:

less

Transcript and Presenter's Notes

Title: Visual Analysis and Semantic Exploration of Urban LIDAR Change Detection


1
Visual Analysis and Semantic Exploration ofUrban
LIDAR Change Detection
  • Thomas Butkiewicz, Remco Chang, Zachary Wartell,
    and William Ribarsky

2
  • Project Goals
  • Develop change detection algorithm
  • Urban areas (new buildings, construction, etc)
  • County-wide, annual LIDAR scans
  • Interactive exploration

Target Users
  • Urban planners
  • Historians
  • Tax Enforcement

3
Airborne LIDAR Laser Rangefinder
GPS/Inertial Navigation Sample Points (x,y,z) w
or w/o classification Tens of thousands of
point/second
4
Previous Work
Many previous approaches interpolate to rasters
Interpolation
3D point cloud
2D grid of heights
Reduces accuracy Points no longer
measurements Fails to exploit areas of higher
density
Vu T., Matsuoka M., Yamazaki F. Lidar-based
change detection of buildings in dense urban
areas. In Proceedings of IEEE Geoscience and
Remote Sensing Symposium, 2004. vol. 5, pp.
34133416.
5
Previous Work
Grids/Rasters than subtracted
Scan 1 - Scan 2
Difference
Change Extraction done on the 2D grid of height
differences
Thresholding
Vu T., Matsuoka M., Yamazaki F. Lidar-based
change detection of buildings in dense urban
areas. In Proceedings of IEEE Geoscience and
Remote Sensing Symposium, 2004. vol. 5, pp.
34133416.
6
Previous Work
Filtering of changes done with image processing
Original Opened
Opened then Closed
Coarse control of filtering due to granularity of
kernel sizes Creates false silhouettes/footprints
Loss of detail
3x3 5x5 7x7
7
Change Detection
Differences in our technique Points tested
individually (No interpolation) Accuracy
preserved Variable density exploited Reliable
measurements 3D, shape-based filtering Finer
granularity More options (size, height, area,
shape, semantical)
8
Change Detection
Input X,Y,Z points Decision Does this point
fit the existing model? Change due to
measurement? or Change in physical
environment? Output Only the X,Y,Z points that
represent changes (To change aggregator)
9
Change Detection
All possible surfaces
Sampled Points
Triangle Network
"Analyzing Sampled Terrain Volumetrically with
Regard to Error and Geologic Variation Thomas
Butkiewicz, Remco Chang, Zachary Wartell, William
Ribarsky Proc. SPIE Visualization and Data
Analysis 2007, San Jose, CA
10
Change Aggregation
Input X,Y,Z Change points (From change
detector) Task Attempt to assemble nearby
points into cohesive models. Output Change
Models (Exported for the Interactive
Application)
11
Change Aggregation
Unvisited marked (as changed) vertices Has
marked neighobrs? No discard (lone point ?
street sign, power line, etc) Yes Add
incident faces to model, visit neighbors
12
Change Aggregation
Input Change Points Output Change Models
13
Interactive Exploration
Raster based methods Results are 2D images
This is boring and offers no insight or
interactivity!
Murakami H., Nakagawa K., Hasegawa H., Shibata
T., Iwanami E. Change detection of buildings
using an airborne laser scanner. Journal of
Photogrammetry and Remote Sensing 54 (July 1999),
148152.
Vu T., Matsuoka M., Yamazaki F. Lidar-based
change detection of buildings in dense urban
areas. In Proceedings of IEEE Geoscience and
Remote Sensing Symposium, 2004. vol. 5, pp.
34133416.
14
Interactive Exploration
Features
3D GIS Environment GIS Database
Integration Interactive Filtering
Tools Analytical Tools
15
Interactive Exploration
16
Interactive Exploration
17
GIS Integration
Other Data Building permits, Tax Database, etc
Vector Data Roads, Building Footprints, etc
18
Non-Realistic Level-of-Detail
Traditionally Popping or fade out Zooming out
Buildings ? 1 pixel ? disappear NR LOD
Solution No popping Seamless transition
Cognitively correct presentation What
abstraction makes sense to show at each
extent?
Distance away
Regions of Changes
Groups of Changes
Individual Changes
19
Full Detail Level
20
Development Level
21
Development Level
22
Regional Level
Legible Simplification of Textured Urban Models
Remco Chang, Thomas Butkiewicz, Caroline
Ziemkiewicz, Zachary Wartell, Nancy Pollard,
William Ribarsky IEEE Computer Graphics and
Applications (CGA) Issue on Procedural Methods
for Urban Modeling.
23
Heat Map Filtering
24
Heat Map Filtering
25
Heat Map Filtering
26
Discussion and Future Work
Additional uses FEMA floodplain mapping Live
battlefield change detection Future
enhancements LIDAR classification data Aerial
photos obliques SAR (penetrability) Automatic
target recognition
Visual Analysis for Live LIDAR Battlefield
Change Detection. Thomas Butkiewicz, Remco
Chang, Zachary Wartell, William Ribarsky. SPIE
Defense and Security Symposium 2008.
27
Conclusions
We have developed A method for comparing LIDAR
data that Preserves accuracy Results in useful
3D change models An interactive application for
exploring the detected changes, which
has Intuitive presentation of
changes Non-realistic LOD for
scale-appropriate data abstraction Analytical
and Filtering tools
28
Questions?
Write a Comment
User Comments (0)
About PowerShow.com