A basic value-range query on remotely-sensed earth science data is constructed in order to find areas over which the geophysical parameter values fall into a given range. An advanced version can have an area size constraint on the results, which will select only those large enough congregated regions as final results. A ternary technique with a two-level data pyramid model has been proposed by our team at GMU in order to speed up answering such queries with an approximate accuracy limit and an area size constraint. A prototype system has been built to illustrate the idea of this technique using NASA's CIDC NDVI data set. On the server side, Java programs and the Splus software are combined together to support index building, searching and convex hull generation. On the client side, SVG (Scalable Vector Graphics) is merged to Java to produce a clear and user-friendly interface. After a user submits a value-range query from the client side, the server searches the data pyramid first and then the actual data. Important stepwise results, such as the derived lower resolution area, the actual data points, and the generated convex hulls, are sent to and displayed at the client side.
[1]
Kwang-Su Yang,et al.
A feasible method to find areas with constraints using hierarchical depth-first clustering
,
2001,
Proceedings Thirteenth International Conference on Scientific and Statistical Database Management. SSDBM 2001.
[2]
Sharad Mehrotra,et al.
Progressive approximate aggregate queries with a multi-resolution tree structure
,
2001,
SIGMOD '01.
[3]
Joseph O'Rourke,et al.
Computational Geometry in C.
,
1995
.
[4]
David A. Duce,et al.
Scalable Vector Graphics SVG 1.0 Specification
,
2000
.
[5]
Michael D. King,et al.
Earth Science Data for All: EOS and the EOS Data and Information System
,
1994
.
[6]
Patrick Brézillon,et al.
Lecture Notes in Artificial Intelligence
,
1999
.
[7]
William N. Venables,et al.
S Programming
,
2000
.
[8]
Anil K. Jain,et al.
Data clustering: a review
,
1999,
CSUR.
[9]
Xiaoyang Sean Wang,et al.
Value Range Queries on Earth Science Data via Histogram Clustering
,
2000,
TSDM.