Automatic generation of Earth Observation Products

The CSIR Satellite Applications Centre (Council of Scientific and Industrial Research) was approached by the South African Government to assist with an implementation plan for the South African Earth Observation Strategy (SAEOS). The goal of the SAEOS is to provide a seamless exchange of information among government departments by providing a warehouse that promotes accessibility to spatial information for all tiers of government.

Since the 1970s, the CSIR has been receiving telemetry from earth observation satellites, building up an archive of remote sensing data going back over 30 years. This archive provides a hugely valuable dataset for the analysis and development of change detection applications. The automation of national information layers derived from remote sensing imagery is created within the center and disseminated to the government.

The portfolio of products will include:

* Urban growth / urban change 1980 – 2007
Forest plantations 1970 – 2007
Buildings / informal settlements
ROP buildings

Customers for these products will be as diverse as the Department of Water Affairs and Forestry, the Department of Agriculture, the Department of Defense, and the South African Police Service. The CSIR has been working with key technology partners to design and implement next generation architecture to meet the goals of the SAEOS by leveraging the existing operational experience in managing the entire remote sensing supply chain to a geo-processed product for analysis by end users.

Solution Architecture

Being a large data-serving entity, the center has a standardized and efficient processing chain to generate and manipulate data into products. This workflow requires the integration of best-in-class technologies to implement key elements of the processing chain. Oracle 10g provides the core repository for geospatial data while the Definiens Enterprise Image IntelligenceTM Suite provides automated quantification and classification from source imagery to automate the generation of new product versions. Each selected technology provides open interfaces and enterprise scalability in order to automate the processing of up to 100 new images per day and create new products from the existing 70TB archive.


Satellite telemetry is received at the center through direct broadcast using the X band antennas. Satellite imagery is converted from telemetry, with the imagery stored in various raw computer-compatible formats and archived in the center. This data is then processed to create the information products, and made available to end users through either a data extract, data base replication or via the internet. The catalogue at the center will be transferred to GLOVIS, an open source Google Earth-type applet used and developed by the U.S. Geological Survey.

An essential element of the solution will be the automation of image analysis, enabling much more efficient turnaround of updates and the generation of new products from the existing archive. The Definiens eCognition® Server provides the batch processing capabilities for image analysis. The image analysis capabilities of the Definiens eCognition® Server is integrated into the overall workflow using the Automation API, designed to provide access to image analysis processing through web services or script-based interfaces. Open input and output connectivity enable a wide range of image source data to be processed while quantification and classification results can be delivered in the appropriate formats to complete the processing workflow.

Crucially, new image analysis applications can be created using the Definiens Developer client and these applications can be further calibrated to prevailing image conditions before execution using the Definiens Architect.

The processing capabilities use the power of the underlying Definiens Cognition Network Technology® to enable the accurate quantification and classification of images, allowing the CSIR to extract deeper insights, generate faster results, and give its customers the ability to make better decisions on the basis of the geospatial information products provided.


The SAEOS is coordinating the collection, assimilation, and dissemination of Earth Observation products into an automated and integrated environment. By integrating the whole process into suitable enterprise architecture, government departments will gain seamless access to key decision-making data in a timely and consistent fashion. Best-of-class technologies like the Definiens Enterprise Image Intelligence™ Suite and Oracle 10g are at the heart of the solution to support the SAEOS, automating and thus further supporting the growth and sustainable development of the South African nation.


Author: EARSC

This website uses cookies to collect analytical data to enhance your browsing experience. Please accept our cookies or read our Privacy policy.