The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008 
203 
A more complex task could be that the user wants to find out all 
images (maybe hundreds or more) inside a specific region and 
then generate full pyramids for each of the images. The 
following simple PL/SQL block would do the work 
automatically. 
declare 
type curtype is ref cursor; 
my_cursor curtype; 
stmt varchar2(1000); 
id number; 
gr sdo_georaster; 
gm sdo_geometry; 
begin 
— 1. define the query area in WGS84 coordinate system 
gm := sdo_geometry(2003, 8307, null, 
sdo_elem_info_array( 1,1003,3), 
sdo_ordinate_array(5,6,30,30)); 
— 2. define the query statement on the georaster table 
stmt := 'select id, t.myimage from my table t' || 
'where sdo_inside(t.my_image.spatialextent, :1)-'TRUE'"; 
— 3. spatially query all images INSIDE the query area 
— and generate full pyramids for each of the images 
open my_cursor for stmt using gm; 
loop 
fetch my_cursor into id, gr; 
exit when my_cursor%NOTFOUND; 
sdo_geor.generatePyramid(gr, 'resampling=bilinear'); 
execute immediate 'update my_table set my_image=:l 
where id=:2' using gr, id; 
commit; 
end loop; 
close my cursor; 
end; 
computing systems. With a multi-tier architecture and the 
power of GRID computing, concurrent processing and 
parallelization can be readily available for raster image database 
management and processing. 
7. CONCLUSIONS 
In summary, this enterprise database-centric approach provides 
a foundation to help solve the two major challenges in a truly 
secure, scalable and performant way and offers an easy-to-use 
interface to empower non-geospatial professionals to manage 
and process geospatial raster datasets. The implementation of 
Oracle GeoRaster based on this database-centric approach and 
the tests we conducted show that this database-centric approach 
is a viable solution for geospatial image management and 
processing. 
This approach focuses on the database server, in which the 
future directions include content-based indexing and search, 
componentizing the server-side image processing and query 
engine and storing them as database models, enhancing raster 
data analysis and mining, leveraging computing clusters and 
parallelizing image processing operations. It is not to discount 
the middleware image servers and desktop image processing 
systems or GIS systems. Instead the spatially enabled database 
server, middleware application servers and desktop image 
processing and GIS systems should complement each other and 
built on top of them a distributed system with a multi-tier 
architecture is the right direction. 
ACKNOWLEDGEMENTS 
The authors would like to thank Zhun Li for conducting some 
of the tests presented in this paper. 
REFERENCES 
Users can also wrap up such blocks into a PL/SQL procedure 
and store it in the database, then call the stored procedure 
directly. Such features enable users to organize complex 
processes and automate database administration tasks easily. 
This API enables non-geospatial experts understand such 
geospatial databases and the data manipulations and thus 
dramatically help expand the raster data usages in broader areas. 
ESRI, 2005. Raster Data in ArcSDE® 9.1 - An ESRI White 
Paper, http://www.esri.com/library/whitepapers/pdfs/arcsde91 - 
raster.pdf (accessed April 2008) 
ISO, 2001. ISO/IEC 13249-5:2001, Information technology - 
Database languages - SQL Multimedia and Application 
Packages - Part 5: Still Image, International Organization For 
Standardization. 
6. LEVERAGING IT INFRASTRUCTURE 
Raster data processing and manipulations are computationally 
complex and I/O intensive. Single process might not be fast 
enough to meet the real-time archiving and processing 
requirements. Concurrent processing offers one of the best 
answers to such requirements. Tests show that concurrent 
processes drastically improve performance and scalability, 
including GeoRaster data loading, queries and processing (Xie, 
Li and Xu, 2006). 
Since GeoRaster allows users to store raster data natively inside 
Oracle databases, it enables all benefits from Enterprise 
Computing technologies, such as the Oracle Enterprise GRID. 
From this standpoint of view, GeoRaster is an Enterprise GRID 
Computing enabler for the geospatial and geoimaging 
applications. Oracle enterprise GRID computing technology 
provides the benefits of lower cost, higher quality and flexibility, 
greater scalability and performance, and so on. In addition, the 
server-side processing engine, as an integral part of the database 
server, can be coupled with middleware and client processing 
systems to fully leverage the power of enterprise distributed 
Oracle, 2004. Oracle Spatial GeoRaster, lOg Release 1 (10.1). 
Oracle Corporation. 
Baumann, ?., 2001. Web-enabled Raster GIS Services for Large 
Image and Map Databases, 5th Int’l Workshop on Query 
Processing and Multimedia Issues in Distributed Systems 
(QPMIDS'2001), Munich, Germany, September 3-4, 2001 
Xie, J., Z. Li, and W. Xu, 2006. Using Enterprise Grid 
Computing Technologies to Manage Large-Scale Geoimage 
And Raster Databases. In: the Proceedings of ASPRS 2006 
Annual Conference, Reno, Nevada, May 1 - 5, 2006. 
Xie, Q., J. Sharma, and J. Ihm, 2007. Oracle Spatial 11 g 
GeoRaster, An Oracle Technical White Paper. 
http://www.oracle.eom/technology/products/spatial/pdf/l 1 gcol 
lateral/spatial 1 lg_georaster_twp.pdf (accessed April 2008) 
Xie, Q., 2008. Oracle Spatial, Raster Data. Encyclopedia of GIS, 
Shashi Shekhar and Hui Xiong (editors), Springer, pp. 826 - 
832.