ul 2004
ient and
e image
memory
he pool.
| all the
chart of
dule
iteria items
ality of data
showed that
high quality
ing able 10
. The image
je viewers
ip between
ternational Archives of the Photogrammetry, Remote Sensin
1 ) |
windows size and transferred data amount based on the three
different operations.
Table 5.1. Windows Size .vs. Data volume
Zoom toa
pyramid
level
Zoom toa beyond pans
Windows neighboring | neighborin | across
size pyramid g level screen
level (KB) (KB) (KB)
640X480 337.84 422.30 370.50
512X380 220.79 283.49 257.07
400X300 150.99 188.73 179.31
320X240 105.46 131.83 130.81
250X190 73.19 91.48 95.38
200X150 52.88 66.11 72.35
The data volume in Table5.1 is the data which will be
transferred to client side to show the highest quality of
images according to users’ requests. The data volume
transferred to client was reduced while the windows size
became small. It shows that the windows size should be
smaller on low network bandwidth than that on high network
bandwidth. The data volume arranges from 52.88KB to
422KB on different windows size and different operations. If
the mean connection speed of Internet users is IMbit/s,
[25KB/s, the sharpening (get all highest resolution images in
that request) time will be from 0.4 to 3.4 seconds. Due to the
data is been streamed to the users, even for the modem users,
they will feel they are getting serves any time after clicking
on the viewer.
6. Conclusions
Based on this improved structure, GSN client becomes much
more efficient and smart than the old one. Firstly, the
response time is equal to a desk top system. User will not feel
any delay while operating map. Secondly, the network will
not be the decisive factor to the end users. Thirdly, many
fancy functions are able to be integrated to GSN client. They
are Fly In/Out, 2D Image Fly (any direction) of 2D Images.
The last feature will be the most important core component
in this project. The module can be easily migrated to a true
3D model if the digital terrain data is available. Finally, it
becomes more reliable and extensible. Further works should
be to use this technique on 3D Presentation and the storage
technologies to handle huge imagery.
By applying streaming and pyramid technologies, the
Sstem's performance has been improved efficiently and
looks like a desktop GIS application. Additionally, enable to
Support GSN Globe has promoted it to be a huge image data
Set server. GSN Image System technology offers these key
features and benefits:
Easy Implementation
GSN Image requires no special server-side software or
complicated authoring tools, allowing for quick and
easy content creation and implementation.
g and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
1143
* . Bandwidth Friendly
GSN Image only downloads the files being used and
streams required files on demand allowing for high-
resolution images over all bandwidth environments.
* High Resolution Images
GSN Image allows web designers the ability to post
high-resolution JPEG and vector graphics files that
users can view, regardless of connection speed.
* ‘unlimited’ amount of images support
If hardware is available, unlimited images can be put to
its storage.
7. References
Harder, C. (1998). Serving maps on the Internet.
Redlands, CA: ESRI Press.
A primer on Internet-based mapping technologies, for
novices. Thomas Baker, Ph.D. PathFinder Science
Federal Communications Commission, 2001.
“FEDERAL COMMUNICATIONS COMMISSION
RELEASES DATA ON HIGH-SPEED SERVICES
FOR INTERNET ACCESS”, US
http://ftp.fec.gov/Bureaus/Common_ Carrier/News Rele
ases/2001/nrcc0133.html
ERMapper White Paper, 2003. "Using ECW
Connector™ and Image Web Server™ with ArcIMS®”,
US.
http://Awww.ermapper.com/document/doc.aspx?doc id=
34
William B. Pennebaker and Joan L. Mitchell, 1992.
JPEG: Still Image Data Compression Standard, Van
Nostrand Reinhold, New York.
J.M. Ogden, E.H. Adelson, J R. Bergen, P.J. Burt. 1985.
Pyramid-based computer graphics, RCA Engineer 30-5
Computer Graphics Group, Remote 3D Visualization
using — Image-Streaming Techniques, Universit at
Erlangen-N"urnberg, Germany http://www.vis.uni-
stuttgart. de/ger/research/pub/pub1999/1$1M A DE99.pdf
8. Acknowledgements
We would like to thank the Geospatial Information and
Communication Technology (GeoICT) Lab, Department of
Earth and Space Science and Engineering, York University
for their support and for providing the GSN source code and
some high resolution images to support the image-streaming
algorithm.
