Spatiotemporal Tile Indexing Scheme Oscar Pérez Cruz Polytechnic University of Puerto Rico Mentor: Dr. Ranga Raju Vatsavai Computational Sciences and Engineering Division August 2009

2Managed by UT-Battelle for the U.S. Department of Energy Overview Problem background Research objective Research background System architecture Application prototype Conclusion and future direction Reference Acknowledgments

3Managed by UT-Battelle for the U.S. Department of Energy Problem background Cause –Past decade witnessed an increasing number of operational satellites –Satellites bring Spectral coverage Spatial coverage Temporal coverage Effect –Amount of data increases through the years exponentially –Difficulty increased in meta data management, and data discovery Needs

4Managed by UT-Battelle for the U.S. Department of Energy Research objective Create a system that can read meta data from satellite images (or ancillary data) to store in a spatiotemporal database that generates –Efficient search and data discovery –Easier access to the meta data –Faster results –Improves data management Current application for the system –Geographic satellite images from  Moderate-Resolution Imaging Spectroradiometer (MODIS)  Advanced Wide Field Sensor (AWIFS) Fig: Africa image taken from the Web Fig: Magnified satellite image taken from the Web

5Managed by UT-Battelle for the U.S. Department of Energy Database is large collection of interrelated data –One database record = one object –Spatial database represent objects in space with a location –Temporal database represent objects in time –Spatiotemporal database represent objects in both time and space Spatiotemporal object have –Identifier –Spatial attribute –Timestamp Data types of a spatial attribute are –Point represents location of entities –Line represents networks –Region represents entities with large areas Spatiotemporal databases

6Managed by UT-Battelle for the U.S. Department of Energy Index structure Ensures fast access to records in a database Based on a search key Avoid a sequential scan through the database Several structures are –Tile devised to index a boundary spatial attribute –Geometric devised to index the spatial attribute –Temporal devised to index time attributes –Spatiotemporal index the spatial and time attributes of an objects

7Managed by UT-Battelle for the U.S. Department of Energy Index behavior on databases Geometric R-Tree Temporal B-Tree Searching for objects in Region a in the year Result: Only object 2 satisfies the parameters of the search query.

8Managed by UT-Battelle for the U.S. Department of Energy System architecture System components: –Meta data harvesting system –Spatiotemporal database  Database management application –PostgreSQL database server  Extension: PostGIS –Application server (PHP) –Web server (Apache) –Web page

9Managed by UT-Battelle for the U.S. Department of Energy Meta data - data that provides information about data Harvesting - process of extracting meta data from ancillary data sources Features: –Uses GDAL open source library  Supports almost 100 image formats  Not dependent of any particular satellite or sensor Meta data elements –Pixel size –Image size –Central meridian –Latitude of origin –Last modified date Meta data harvesting system Fig: Meta data harvesting system model

10Managed by UT-Battelle for the U.S. Department of Energy Spatiotemporal database management application Features: –Connect to a database –Create, delete, view, and index tables –Harvest meta data to store in database –Search for stored meta data of images Fig: Spatiotemporal database management application menu model

11Managed by UT-Battelle for the U.S. Department of Energy Web page Advantages –Global access –Easy to use –Only one application Features –Option for search parameters  Date  Location  Date and location –Format for location parameter  DMS format  Decimal degrees format Fig: Web page and application server models

12Managed by UT-Battelle for the U.S. Department of Energy Querying database Spatial Query retrieves all objects whose geometry contains a given point. SELECT * FROM WHERE ST_WITHIN(‘POINT(3 2)’, LOCATION); Temporal Query retrieves all objects whose time has to do with an event. SELECT * FROM WHERE DATE = “12/12/2000”; Spatiotemporal Query retrieves all objects whose time has to do with an event and geometry contains a given point. SELECT * FROM WHERE ST_WITHIN(‘POINT(4 2)’, LOCATION) AND DATE = “12/12/2000”;

13Managed by UT-Battelle for the U.S. Department of Energy Other query examples Spatial Query that returns regions objects whose distance of separation is within the parameter. SELECT * FROM WHERE ST_DWITHIN(‘REGION (geometric_information)’, LOCATION, distance); Spatial Query that returns regions objects who overlap each other. SELECT * FROM WHERE ST_OVERLAPS(‘REGION (geometric_information)’, LOCATION); Spatial Query that returns regions objects that are within other region objects. SELECT * FROM WHERE ST_WITHIN(‘REGION (geometric_information)’, LOCATION);

14Managed by UT-Battelle for the U.S. Department of Energy Database management application prototype Menu (0) Connect to the Database (5) Make an Index for the Table (3) Delete table (1) Create a new table (4) Insert meta data of an image on a table

15Managed by UT-Battelle for the U.S. Department of Energy Web page prototype

16Managed by UT-Battelle for the U.S. Department of Energy Conclusion and future direction System enables easier data management of satellite images Current stage of system development –Database development –System components upgrades Future extension for the system –Other kinds of search criteria like:  Cloud coverage percentage  Forestation percentage  Population growth  Type of resolution –Interface other data formats –Meta data search through the network

17Managed by UT-Battelle for the U.S. Department of Energy Reference Hartmut Güting, R., and Schneider, M. (2005). Moving Objects Databases. San Francisco: Morgan Kaufmann Publishers. Rigaux, P., Scholl, M., and Voisard, A. (2002). Spatial Databases With Application To GIS. San Francisco: Morgan Kaufmann Publishers. PostgreSQL Documentation, The PostgreSQL Global Development Group. Neufeld, K. (2009). PostGIS 1.4.0rc2 Manual. Geospatial Data Abstraction Layer (GDAL), GDAL/OGR Project Management Committee,

18Managed by UT-Battelle for the U.S. Department of Energy Acknowledgements Debbie McCoy (RAMS Program Manager) GIST Group Office of Advanced Scientific Computing Oak Ridge National Laboratory U.S. Department of Energy

19Managed by UT-Battelle for the U.S. Department of Energy Questions