Mobile Device Visualization of Cloud
Generated Terrain Viewsheds
Chris Mangold
College of Earth and Mineral Science
Penn State University
State College, PA
csm202@psu.edu
Advisor: Dr. Peter Guth
Motivations
 Mobile visualization of GIS data
 Products of Terrain DTM/DSM spatial analysis
 Cloud GIS
 Mobile
 Augmented Reality (AR)
Rothera Point, Adelaide Island, Antarctica. Aster (v2) Global DEM overlay.
Augmented Reality (AR) in GIS
Libertytown, MD (layar,2014)
Yelp urban guide (Yelp,2014)
 Location Intelligence (LI) Mobile Apps
 Point vector based
 AR frameworks
 Next Generation
 3-D model rendering
 Raster data based
Fai della Paganella Trento, Italy
(Dalla Mura, 2012)
Least Observed Path (LOP) Application Concept
LI Mobile Application
 Provides a navigation path to avoid detection
 Renders AR geo-layer
 Consumes Cloud generated observer viewsheds
Cloud hosted GIS
LOP System Diagram - Work Flow
 Define LOP environment
 Request and consum observer viewshed results
 Geo-register result using devices sensors
 Generate and render AR geo-layer
Cloud GIS
2 KM Radius RF Propagation IFSAR 5 M
2.5 KM Slope Position Classification
IFSAR 5 M
1.7 KM Observer Viewshed IFSAR 5 M
(MrGeo, DigitalGlobe 2014)
 Computing Efficiencies
 Apache Hadoop MapReduce framework
 Virtualized commodity and clustered resources (GPUs)
 Terrain spatial analysis web services
 REST APIs
LOP Application UI
(Map View – Device Horizontal Orientation)
Map View
 OSMAnd open source framework
 Slippy map user interface
 Drop pin to identify observer locations
 WGS84 Web Mercator MBTiled base map
LOP Application UI
(Augmented Curtain View – Device Vertical Orientation)
Augmented Curtain View
 Renders AR curtain layer
 Recalculated as device location updates
 POSE derived from orientation sensors
 Visibility probability color ramp indicator
NED 1”
NED 1/3”
Lidar 10 M Aggregate Generalization
Lidar 3M Aggregate Generalization
Data source
Elevation model
ASTER GDEM 1”(~30 meter resolution)
DSM
Lidar 1 meter
DSM
NED 1” (~30 meter resolution)
DTM
NED 1/3” (~10 meter resolution)
DTM
SRTM 3” (~90 meter resolution)
DSM
Lidar – 1.0 Meter
LOP Augmented Curtain Generation
AOI curtain base evaluation image
Scale: 1 Pixel = 1 Meter
Scale received viewshed PNG images
Geo-register and merge images
Create evaluation bitmap
 Size bitmap to LOP evaluation AOI
 Normalize and scale viewshed images
 Geo-register images
 Merge and clip images to AOI
LOP Augmented Curtain Generation
Create AR curtain base
 Array of 360 RGB values
 Evaluate pixels within AOI
 RGB values to determine
visibility
 Calculate azimuth to location
 Track total and visible pixel
Visualization of calculated AOI curtain base.
 Calculate azimuth weighted value
LOP Augmented Curtain Generation
Render LOP geo-layer
 Overlay on Android surface view
 Determine screen orientation and size
 Apply weighted visibility for each azimuth
 Draw compass components
Augmented Curtain POSE
POSE
 AR: integrating virtual data with real world
 Enhance geo-register LOP curtain layer
 Manage device inertia sensors
 Magnetic
 Gravity
 Kalman filter
 Smoother rendering
LOP Application Evaluation
LOP evaluation site.
LOP site looking north through alley.
Environment
 Suburban office park setting
 Droid Incredible
 Target observation height 2 meters
 LOP AOI 200 m diameter
Viewshed origin point looking west.
LOP Application Evaluation
LOP basemap with viewshed overlay.
Measure
 Observer viewshed cloud request time
 Time to render LOP augmented curtain
 Detection of a LOP
LOP Application Evaluation
NED1” and other bare earth returns
 Performance response times < 0.5 seconds
 No detected LOP
LOP Application Evaluation
Lidar 10m
 Performance response times < 0.5 seconds
 Contiguous LOP path between 29.0o - 39.0o
LOP Application Evaluation
Lidar 3 m
 Performance response times < 0.5 seconds
 Contiguous LOP path between 34.0o - 40.0o
LOP Application Evaluation
Lidar 1 m
 Performance response times < 0.5 seconds
o
o
 Broad low LOP probability area (25.0 - 45.0 )
 Distinct LOP sections between 26.0o - 37.0o
Conclusions
 LOP, demonstrates geo-visualization of Cloud
generated viewsheds
 Add outlier filtering algorithms for 1 m Lidar
 Small LOP AOIs show no performance penalty
Future directions
 Evaluate LOP with larger spatial extents
 Optimize rendering algorithms
 Add depth projection to LOP curtain
 Investigate edge detection
 Evaluate porting application to Google Glass
Questions
 LOP, demonstrates geo-visualization of terrain based
raster data
 Add outlier filtering algorithms for 1 m Lidar
 Small LOP AOIs show no performance penalty
Sources
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Apache, 2013, ApacheTM Hadoop: http://hadoop.apache.org/, last accessed 19 Feb 2014.
Amazon, 2014, Amazon web services – Amazon EC2: http://aws.amazon.com/ec2/, last accessed 19 Feb 2014.
Baboud, Lionel. M. Cadik, E. Eisenmann and H.P. Seidel, 2011. Automatic photo-to-terrain alignment for the annotation of mountain pictures, IEEE
Conference on Computer Vision and Pattern Recognition, pp. 41-48
Dalla Mura, M., M. Zanin, C.Andreatta and P.Chippendale, 2012a. Augmented reality: Fusing the real and synthetic worlds, IEEE International Geosciences
and Remote Sensing Symposium, (8):170-173.
Dalla Mura, M., and P.Chippendale, 2012b. Real-World DEM Harmonization through Photo Re-Projection, IEEE International Symposium on Geoscience and
Remote Sensing, 2012. (7): pp 428-430
DigitalGlobe, 2014, DigitalGlobe – Defense & Intelligence http://www.digitalglobe.com/Industries/defense%26intelligence#overview, last accessed 17 Feb
2014.
Fielding R.T., 2000. Architectural Styles and the Design of Network-based Software Architectures, University of California, Irvine, 2000
http://www.ics.uci.edu/~fielding/pubs/dissertation/top.htm , last accessed 17 Feb 2014.
GeoTools, 2014, GeoTools - The Open Source Java GIS Toolkit: http://www.geotools.org/, last accessed 17 Feb 2014.
Google, 2014, GLASS: http://www.google.com/glass/start/, last accessed 17 Feb 2014.
JKalman, 2013, JKalman http://sourceforge.net/projects/jkalman/, last accessed 17 Feb 2014.
layar, LAYAR SDK: https://www.layar.com/developers/, last accessed 17 Feb 2014.
Mapbox, 2013a, TileMill: https://www.mapbox.com/tilemill/, last accessed 17 Feb 2014.
Mapbox, 2013, MBTiles tileset format: https://github.com/mapbox/mbtiles-spec, last accessed 19 Feb 2014.
NASA, 2014a, ASTER: http://asterweb.jpl.nasa.gov/, last accessed 19 Feb 2014.
NASA, 2014b, SRTM: http://www2.jpl.nasa.gov/srtm, last accessed 19 Feb 2014.
OsmAnd, 2014, OsmAnd – Map & Navigation: http://osmand.net/ , last accessed 17 Feb 2014.
Noguera, J.M., C. Ogayar and R. Joan-Arinyo, 2013. A scalable architecture for 3D map navigation on mobile devices, Personal and Ubiquitous Computing,
17(7): pp. 1487-1502.
Porzi, L., E. Ricci, TA. Ciarfuglia, and M. Zanin, 2012. Visual-inertial tracking on Android for augmented reality applications, IEEE Workshop on Environmental
Energy and Structural Monitoring Systems,(8):35-41.
Rosenberg, J., and A. Mateos, 2011. The Cloud at Your Service, Manning, Greenwich, pp 146-146.
Yelp, 2014, Yelp http://www.yelp.com, last accessed 17, Feb 2014.
USGS, 2013, The USGS Store – Map Locator & Downloader: https://store.usgs.gov/b2c_usgs/usgs/maplocator, last accessed 20, Dec 2013.
USGS, 2014, National Elevation Dataset: http://ned.usgs.gov/, Last accessed 17, Feb 2014.
Wikitude, Wikitude SDK: http://www.wikitude.com/products/wikitude-sdk/, last accessed 17 Feb 2014.
William and Mary, 2014, Virginia lidar: http://www.wm.edu/as/cga/VALIDAR/ , last accessed 08 Feb 2014.