The Evolution in the Geoprocessing Technology

This page shows a retrospective view and a prospective view about geoprocessing technology, where it is considered the three generations of Geographic Information Systems.

The first generation (“cartographic CAD”) is characterized by a system inheriting from the cartography tradition, with limited database support where the typical working tool is a map (called “coverage” or “information layer”). These systems were designed in the 80s for the VAX environment and - starting in 1985 - for PC/DOS personal computer systems, this system class is mainly used in isolated projects , without being concerned about generating digital data files. This generation also can be characterized as project oriented systems (“project-oriented GIS”).

The second GIS generation (“Geographic database”) arrived in the market at the beginning of the 90s and it is characterized by its client-server environment concept, attached to relational database managers and additional packages for image processing. These systems were designed for multi platform environments (UNIX, OS/2, Windows) with a windows based interface, this generation can also be viewed as a supporting system for institutions (“enterprise-oriented GIS”).

It is possible to forecast, for the end of the 90s, a new GIS generation (“digital geographic libraries” or “Geographic data centers”), managing large databases of Geographic data, with network accessibility, either local or WANs, with a WWW (World Wide Web) interface.

For this third generation, the growing of spatial database and the requirements for sharing them with other institutions requires distributed and federative database technologies. These systems must follow the interoperability requirements, such that it allows access for spatial information for distinct GISs.

The GIS third generation can still be viewed as the system development for information exchange among institutions and the other society components (“society-oriented GIS”).

The next figure shows the evolution of the Geoprocessing technology.


Evolution of the GIS technology.

Introduction to Geoprocessing

The first generation: cartographic CAD

The first GIS generation is characterized by system with graphical and spatial analysis operations on files (“flat files”). Its connection with database managers is only partial (part of the descriptive information are found in the file system) or non-existent.

They are more adequate for spatial analysis project development over small or mid size regions. These systems emphasize mapping aspects. The system allows the data input without previous definition of the conceptual scheme, looking like CAD environments that have cartographic projections representation capability and also attributes and spatial objects association. By conception, these environments do not have adequate support to build large spatial databases.

Utilization
The first and natural GIS usage in the majority of the organizations is as a tool for basic mapping production. Given the absence of adequate information for decision making over urban and environmental problems in Brazil, using the Geoprocessing technology in this matter is already a valuable contribution. Unfortunately, many institutions are careless about keeping a digital database of their results. Consequently, very important results are inaccessible.

To mention just a single example, the “SOS Mata Atlântica” project produced one of the largest worldwide studies using the GIS technology. More than 200 charts in the 1:250.000 scale were produced showing all the coverage of original tropical forest remains, using photo-interpretation of satellite images. Besides the excellent quality obtained, the final results were not consolidated in a Geographic database.


Introduction to Geoprocessing


The second generation: Geographic Databases

The second GIS generation is characterized by systems conceived to operate as a Geographic database. It was designed as a non conventional database where the data handled has descriptive attributes but also a geometrical representation in the Geographic space.

The second system generation is still under development, with new products being designed and introduced in the market. No available product today shows all requirements presented here although the industry have indicated their product releases should be compatible with the following requirements.

REQUIREMENTS
Extending the database technology to work with Geographic data requires a progress in several aspects, such as:

  • Conceptual modeling
  • Progress in the Geoprocessing conceptual modeling are required to break the raster-vector dichotomy (NCGIA, 1989) and generate interfaces with larger semantic contents (Egenhofer, 1989).
  • Remote Sensing - Geoprocessing Integration
  • For environmental applications, one of the most important requirements for spatial analysis is the integration among thematic maps, terrain models and satellite images (Ehlers et al., 1989).
  • Additional motivation comes from the recent progress in the Satellite Image Processing area, using new types of sensors (synthetic aperture radar ) and new automatic image interpretation techniques (image segmentation using neural networks).
  • Topological representations in Multi scales and projections.
  • The management of a large Geographic database system requires that multiple geometrical representation associated to the same Geographic data are kept by the system. Consider, for instance, in the different representations associated to the Amazon river in a database covering all the region but also Geographically broken according to the UTM zone. In such case, the system has to separate the data representation and also guarantees the unique description.
  • Consulting, Manipulation and Presentation Language
  • The second GIS generation requires consulting, manipulation and spatial object representation languages with a high expressive power. The consulting language can be based on SQL (as presented in Egenhofer, 1994), the Geographic data manipulation follows the same line as the map algebra (Tomlin, 1990; Câmara, Freitas e Cordeiro, 1994).
  • New Geographic Analysis Techniques
  • The incorporation of new techniques for Geographic analysis such as continuous classification (Druck and Braga, 1995) and environmental modeling (Kemp, 1993) is required so the GISs are capable of satisfying the requirements for analysis and modeling in large spatial databases.
  • Large Database Architectures
  • The complexity and the size of Geographic data requires changes in the traditional architecture schemes in database management systems. Besides the progress in architectures (such as technologies related to distributed systems) it is required to have new spatial indexing methods, which are adequate to the large amount of data to be managed.

Utilization
Although the second GIS generation systems have introduced many benefits for Geoprocessing applications, their usage in Brazil have been very limited. The main reasons are of institutional order.

The client-server environment usage requires competence in Database administration and Computer Networks, which is not available in institutions that use Geoprocessing. It also requires a larger investment to buy, install, and operate database management systems (DBMS) in the market. Additionally, the large integrator potential of GISs can only be used when the system is integrated with production process. This requires that corporative databases have to be translated to the same DBMS environment used by the GIS.

An excellent example of the GIS usage in the client-server environment is the SAGRE system, developed by CpqD/TELEBRAS. Starting from the support offered by the GIS (VISION) and by a DBMS with long fields (ORACLE), it was built facilities for operation and management of the telephone network. This environment is being installed in telephone companies all around Brazil.


Introduction to Geoprocessing


The third GIS generation: Digital Geographic Libraries

A digital Geographic library (or a “Geographic data center”) is a Geographic database shared by a set of institutions. This library has to be remotely accessible and stores, besides the Geographic data, description about the data (“metadata”) and associated multimedia documents (text, photos, audio, and videos).

This new paradigm is motivated by our increased interest on the ecological, urban, and environmental problems , trying to understand, in details, local and global changing processes, and by the necessity to share data among the institutions and the society.

The basic core of a digital Geographic library is a large Geographic database. To show it, imagine two scenarios: a large Brazilian environmental center and a planning secretary for a mid size city, say about 1 million inhabitants: ·

  • First Scenario (“Amazon Database”): This database would have detailed information about human occupation in the region, including basic themes (vegetation, pedology, geomorphology), human occupation (deforesting, farming and cattle raising projects, mineral prospection areas), derived themes (economical zones) and updated satellite images. This database could be kept by a central institution, say the IBAMA, allowing concurrently access for researchers all around the country.
  • Second Scenario (“Curitiba City Hall - city planning office”): This database would have all required information about the city main plan, including: lots, blocks, streets, services (hospitals, schools), water and sewerage system, electrical network. It could be accessed on-line by several city offices, by companies, or even by citizens.

In these two scenarios, the environment has to guarantee concurrent access to a user community, with different selection methods, including browsing and consulting languages.

An example of this technology is the EOSDIS (EOS Data and Information System) system, designed by NASA to be used with the EOS (“Earth Observation System”) program, including a set of satellites with large capability for data generation (2 Tb/day). Considering the general view, the following sections present the main requirements for this new GIS generation, divided into large areas: metadata availability, Internet access, browsing and interoperability.

Requirements: Metadata
The metadata idea (or “data over data”) is to create an environment that presents general descriptions about the data sets locally available or stored in associated centers. An example of such a system is the NASA's MFD “Master File Directory”, including references for data sets in different institutions and space agencies in other countries. One of the main challenges for the construction of metadata environments is to balance the required complexity on describing data collections, because the final information has to be enough to guide the search (Abel and O’Callaghan, 1992). The NASA's MFD adopts the philosophy to keep the minimum set of mandatory descriptors, with fields using free text for additional information; this strategy minimizes the efforts required to make the reports about the available search capacity.

Another desired aspect in the metadata environment is the availability of synthesis data, such as maps in very reduced scales, that could be used to Geographically localize the available data sets.

The ideal situation would be to allow a consulting refinement process, establishing a continuous path between the more abstract metadata level and the data itself (Smith and Frank, 1989). Thus, from the general aspects of the information (“Amazon Database”), it is passed to regional views (“Roraima State”), and specifics (“Roraima Vegetation Maps”) up to the data itself (“The Demene river region Vegetation map in the 1:100.000 scale”).

Requirements: Internet Access
The availability of multimedia interface through the Internet, given by the WWW environment, allows Geographic data to be presented in a pictorial way (using reduced maps and “quick-look” images).

However, the WWW environment (using the HTML - Hyper Text Mark-up Language) presents some problems when using together with Geographic databases, mainly because the HTML is navigational and does not carry transaction capabilities. In order to combine in an appropriate way the WWW/HTML technology with a typical consulting environment for Geographic databases, it is inevitable to use, besides the pictorial resources of HTML, the consulting language with spatial constraints.

Requirements: Pictorial Navigation
The pictorial navigation (Browsing) can be seen as a selection based on pointing: an interactive interface allows the user to go through a database.

It is important to provide navigation mechanisms, because it is not possible that the user previously knows which data types are available and how to proceed in order to get access to the data. The navigation tools allow the user to access data based on its spatial location. The figure below show a possible partitioning by index-maps from a country, which could be used for navigation.


Index-maps structure for navigation.


The great challenge when assembling a navigation environment is to guarantee a quick response and interactivity. To give a quick response, it is required to have generalization mechanisms, which must be different according to the data formats. It can also be required , by performance reasons, a pre processing step, using the required data for the browsing process.

Requirement: Interoperability
Data and procedure sharing among Geographic databases based on distinct GISs is a considerable challenge. As there is no single Geographic data model universally accepted (as opposed to the relational model used in conventional applications), the different GISs in the market show important differences about how they operate, internal storing formats.

When trying to fix these problems, several countries have being working to define cartographical standard for data transfer, keeping the Geographic information richness (topology and attributes). These patterns try to be “neutral” to several companies, including the SDTS (Spatial Data Transfer Standard) in the USA and the SAIF (Spatial Archive and Interchange Format) in Canada.

However, even if the data transfer among Geographic databases is established with different formats, remember that, in complex applications, such as Geoprocessing, a substantial part of the information holds on the consulting, manipulation and presentation procedures. Thus the data interchange does not guarantee interoperability.

In this thesis context, there is no way to explore in higher details this fascinating topic, which we hope can be studied later.


TECHNOLOGY EVOLUTION
Based on Geoprocessing technology evolution presented above, the intention is to act in two different aspects:

  • establish solid concepts for the development and refinement of a second generation GIS (SPRING);
  • start ideas for a project and develop the technology for Geographic data centers, which will complement the SPRING capabilities.

The main purpose of SPRING 2.0 is to be a GIS satisfying all the second generation requirements, that is, a “Geographic Database”, so we believe that complete Geographic database is fundamental in the process of building efficient digital Geographic libraries.


Geoprocessing Evolution


Geoprocessing Introduction