Web-Based Value-Added Satellite Image Services for Agro-Environmental Monitoring

This collaboration with Prof. Bauzer Medeiros at UNICAMP in Brazil aims at establishing a Brazilian/German research cooperation on the integration of large-scale satellite image archives into online information systems. It is a multidisciplinary project that applies IT research to agro-environmental applications.

On the Brazilian side, this project is related with the WebMAPS II project within CNPq`s Universal RFP, whose goal is to solve theoretical and implementation problems concerning scientific data management for agriculture. WebMaps II is part of an ongoing research effort coordinated at the Institute of Computing, UNICAMP, involving CS and Agriculture researchers, as well as several PhD and MSc students. One of the open problems faced concerns construction and organization of large databases containing several Terabytes of satellite images, as well as data collected from sensor networks.

On the German side, proponents have a successful research record of dealing with these kinds of data. Jacobs University Bremen investigates on advanced services for multi-dimensional raster data, such as 2-D image archives, 3-D image time series (x/y/t) and exploration data (x/y/z), and 4-D oceanographic and atmospheric data (x/y/z/t). The technology serving as research platform, rasdaman ("raster data manager"), allows to maintain raster data in standard databases, such as the open-source PostgreSQL. Rasdaman has received several innovation awards and is used in operational services by European mapping agencies and mining industry. Jacobs University is actively participating in the Open GeoSpatial Consortium (OGC) to develop and advance geo raster services.

Research results are being applied to a specific application domain, with social and economic relevance all over the world. This research is associated with two of the five Grand Research Challenges in CS defined by the Brazilian Computer Society (Management of information over massive volumes of distributed multimedia data; Computational modeling of complex systems: artificial, natural, socio-cultural, and human-nature interactions).

Emphasis is on defining strategies for integrating and processing heterogeneous scientific data, collected from several sources, primarily satellites and ground sensors. Results range from low-level storage structures to store these data, to query language constructs and functions to help end users manage these data, making them available on the Web. More specifically, they involve: (1) new techniques, algorithms, methodologies and software for management of massive volumes of scientific spatio-temporal data; and (2) algorithms and methods for reuse and management of digital content in databases, for remote sensing data.

These results contribute to constructing a basic computational platform which can be used by scientists, but also to support formulation of policies in agriculture. This is an area that poses several challenges to computer scientists, and where most solutions are proposed by experts in agriculture, who do not take full advantage of CS research. Further, by deploying an OGC standard currently under development, WCPS, an important contribution to OGC's own quality control is done.

Besides the main researchers on each side, the project involves junior researchers as well as graduate and undergraduate students both in Brazil and Germany. Though the project is motivated and validated by demands of agriculture, the results are expected to be applicable to any kind of domain that demands management of large sets of sequences of image data, such as applications on maritime information systems, biodiversity studies, pollution management or patient monitoring using medical images.

Sponsor: BMBF (Germany), CNPq (Brasil)