199
AN ENTERPRISE DATABASE-CENTRIC APPROACH
FOR GEOSPATIAL IMAGE MANAGEMENT AND PROCESSING
Qingyun (Jeffrey) Xie, Siva Ravada, Weisheng Xu, Zhihai Zhang
Oracle USA, Inc., One Oracle Drive, Nashua, NH 03062, USA -
(qingyun.xie, siva.ravada, weisheng.xu, zhihai.zhang)@oracle.com
KEY WORDS: Spatial Databases, Raster Data, Data Management, Image Processing, Query Processing, Database, Software
ABSTRACT:
Geoimagery and raster gridded data are growing exponentially. As a result, we face many challenges. This paper discusses two of the
major challenges. One is how to effectively and efficiently archive, manage, process and deliver all those geoimage data. Another
challenge is how to make the geoimages and professionally extracted information available to broader audiences so that enterprises
and mass consumers can benefit more from our work. This paper focuses on the database server technology, which is one of the key
areas that is essential and foundational for solving the above two challenges and beyond. It proposes a new enterprise database
centric approach for geospatial image and raster data management and processing. The key concept of this approach is to enhance
and leverage enterprise IT technologies and provide a database-centric solution for image data management as well as key image
processing operations. It consists of three major components: a new native database data type for raster data, a server-side image
processing and raster operation engine and a standard-based user-friendly interface. It’s designed to work in a client-server
environment as well as in any multi-tier architecture. The advantages and benefits of this approach are discussed. The Oracle Spatial
GeoRaster was designed based on this approach. A series of tests and research using the Oracle GeoRaster technology were
conducted and are partially presented in this paper. The results show that this approach is practical, easy-to-use and truly scalable and
performant. This database-centric approach is a viable solution for geospatial image management and processing.
1. INTRODUCTION
Geoimagery and raster gridded data are growing exponentially.
Numerous remote sensors of different types on various
platforms are collecting real time data about the Earth and our
environment for different purposes on a daily basis. As a result,
we face many technical challenges, two of which are the major
ones we need to address carefully. One is how to effectively and
efficiently archive, manage, process and deliver all those
geoimage data. This real-time processing, management and
distribution task is becoming overwhelming and we have to
solve it intelligently. Another challenge is how to make the
geoimages and professionally extracted information available to
broader audiences so that a variety of businesses and mass
consumers can benefit more from our work. In other words,
geospatial and geoimaging technologies need a good platform
to go mainstream. New research and technologies are needed in
order to better solve those problems and the existing database
and application server and client technologies must work
synergistically in order to achieve those goals. This paper
focuses on the database server technology, which is one of the
key areas that is essential and foundational for solving the
above two challenges and beyond.
It’s well known that enterprise database management systems
provide great data security, reliability, availability, recovery and
backup, transactional features, versioning and concurrency, to
name just a few. Because of these benefits, spatial database
technologies based on standard relational database management
systems (RDBMS) have become very popular in recent years.
Good examples include RasDaMan/RasGEO (Baumann, 2001)
and ArcSDE (ESRI, 2005). They typically take a middleware
approach by storing data inside a standard RDBMS system and
processing the data in a middleware or client software package.
Most RDBMS’s don’t have image data types defined. So this
approach requires a relational database schema to be designed
to store the imagery inside RDBMS. However, a fixed set of
relational tables specified in such application schema doesn’t
offer a good flexibility when it comes to integrate geoimage
datasets with other enterprise datasets. The middleware acts as a
processing engine, which queries the data from the RDBMS,
retrieves the data out and then process the data in the
middleware to serve the clients. Some extra data might have to
be retrieved and delivered. Data transportation is expensive and
could be insecure. So performance and data security are
concerns with this approach. The other downside is either the
lack of standard database SQL interface or the decoupling of its
interface from the RDBMS system, which significantly limits
the usability and enterprise integration efforts.
In this paper, we describe an enterprise database-centric
approach for geospatial image and raster data management and
processing. The key concept of this approach is to enhance and
leverage enterprise IT technologies and provide a database
centric solution for image data management as well as key
image processing operations. It not only offers the
aforementioned standard database benefits but also goes one
step further to provide more advantages to tackle the specific
requirements derived from the two major challenges facing the
geoimaging and geospatial industry.
2. THE ENTERPRISE DATABASE-CENTRIC
APPROACH AND THE BENEFITS
While it’s been proven to have many advantages and have
become the industry trend to manage geoimagery data in
commercial RDBMS systems instead of directly using file
systems, we think the traditional RDBMS system itself should
be enhanced to specifically handle geoimagery within the
databases as well.