Technotes-Plain Sailing
Implementing a Digital Nautical Chart Production System for the US
By S.J. Fletcher

Mapping and charting has entered a new era. Historically, the discrete data types in your map database had to be separately updated - the photogrammetric data didn't know what the survey data was doing, the survey data didn't know what the remote sensing data was doing, and so forth. Now, however, you can model, manage and maintain all your geospatial data in one resource, simultaneously - a huge step in ensuring the integrity and accuracy of your mapping and charting media.
    The United States National Imagery and Mapping Agency (NIMA) is presently benefiting from this organic approach to geospatial data management and is currently investigating a vital resource - its Digital Nautical Chart Production System (DNCPS) - around the technology. The way in which NIMA's needs have been addressed by this new generation of solutions illustrates the direction which geospatial technology in general must now take.

Specifying The Solution
The primary focus of the NIMA project was to define and establish the data generation, population, attribution, and maintenance procedures necessary for an active OO production database environment. The DNCPS developed would meet very specific requirements for source data input, data compilation and maintenance, digital product finishing and external system interfaces.
    NIMA's requirement was also that this DNCPS solution be built effectively using existing COTS (commercial off-the-shelf) geospatial products. The supplier chosen was Laser-Scan, based in Sterling, VA, and Cambridge, United Kingdom. The products were VTRAK, currently the market's most powerful raster-to-vector conversion tool, and the Laser-Scan Automated Map Production System (LAMPS2). Both products are in use with mapping and charting agencies worldwide, with especially significant installations at Land Information New Zealand, Ordnance Survey (UK), the South African Navy Hydrographic Office and Mexico's INEGI.
    The entire DNCPS project forms part of a larger NIMA program, Hydro Vision. The aim of this program is to promote the advantages of imagery, imagery intelligence and hydrographic data sources, by providing customers with "up-to-date, tailored data on demand, from a single digital nautical data maintenance environment, focused on aggressive data collection and near real-time information processing by regional teams."
    Hydro Vision, when fully achieved, will provide continually up-to-date nautical information to mariners, when and where it is needed, in a format that can be used effectively. NIMA is moving to an all-digital environment to guarantee its users the information edge.
    NIMA's initial DNCPS production environment consists of several Sun SPARC workstations, running LAMPS2 and VTRAK, which are used for data capture, maintenance and finishing. Initial data sources include CD-ROM and tape, with VPF data. From this, the system produces digital products and digital datasets.
    The initial aim of the DNCPS is to model and manage the databases underlying two Digital Nautical Charts, DNC 8 (the western coast of Europe) and DNC 17 (the East Coast of the United States). Database compilation and output are essential elements in this procedure, as are the inclusion of weekly updates in response to Government-published Notices to Mariners (NTMs) and other information sources. The final dissemination of the charts will take place in a number of media, including tape, CD ROM and the Internet.
    After DNC 8 and 17 have been completed, DNCPS could be used for producing all or part of the remaining 27 of the 29 regional datasets needed for global coverage.

The Underlying Technology - Objective Benefits
Before NIMA's use of the technology can be fully understood, it is necessary to look at the unique features of the technology itself. These can be summarily expressed in one concept - the Gothic geospatial database. Gothic is contained within LAMPS2 and enables NIMA to accurately model - rather than just depict - the world. We have already briefly mentioned the value of simultaneously managing different types of data in one central resource, in order to ensure that updates are consistent and that data integrity is guaranteed. Gothic is an extremely effective way of doing this, because of its special features.
    Firstly, rather than abstracting the world as lines, nodes and three-dimensional strings, Gothic uses an object-oriented (OO) approach; that is, it treats the world as the collection of objects (features) which it really is. In nautical terms, these objects would include buoys, cliffs, icebergs, shipping lanes, and so on, but the definitions might equally be extended to cover objects in land cartography (roads, houses, trees) or other non-geographic data (consumer spending patterns, laser reflection models - to name but two object types in use within Gothic elsewhere in the world).
    Secondly, Gothic stores this intelligence in the database itself, instead of in the application software (which is where other systems usually store it). Consequently, the intelligence can be "plugged into" numerous other types of spatial applications - aeronautical, topographic, photogrammetric, business analysis, network modeling and so on. The lateral uses for the technology are exceedingly diverse - but each application makes use of a common analytical intelligence, in the Gothic core.
    This ability to share intelligent data between applications is an obviously cost-effective method of helping end-users to realize the potential of their data across a number of areas of endeavor. Specifically, the system's absolute integrity across multiple data types is of particular relevance to NIMA, as the DNCPS will also incorporate imagery and intelligence data, in addition to the secondary cartographic data sources mentioned earlier.

Pinning The Features Together
A key phrase which NIMA has used in connection with the DNCPS project is "the stake through the database." What is meant by this, is that in a complex system such as NIMA's, which contains multiple different feature and object libraries at multiple scales, and for a variety of different nautical uses and "coverages," there needs to be some way of linking or "pinning" features and objects to each other through each and every layer. This is so that they update uniformly and behave appropriately in whatever context they are represented. Essentially, features should be "extracted once, maintained many times."
    This task is not without its difficulties. Traditionally, smaller-scale charts (such as Approach) have been collected independently of larger-scale charts (such as Harbor) in the same area. The two classes of chart may draw on each other's data, but there has been no required linkage between them.
    One way of forming this link is by creating a "master feature layer," which helps to establish identity between the same features represented at different scales and in different coverages. Changes applied to one guise of a feature automatically become changes applied to that feature in all its guises. LAMPS2 automatically alters geometry according to scale (this process is known as generalization) and thus considerably accelerates the creation of the master feature layer.
    Likewise, the master feature layer must take into account the fact that different features can participate in different types of coverage. A particular bridge feature, for example, could appear in a coverage called "Obstructions to navigation" as well as in a coverage called "Cultural," both of which in turn could form part of a coverage called "Transportation."
    The object classification system within Gothic fully enables representation rules of this kind to be built into the master feature layer. NIMA's feature classes within the DNCPS number eleven; Gothic can comfortably handle thousands of such classes, each containing millions of features, with total integrity.

Conclusion:
Making The Most Of Your Maps There has not been the time or the space here to explore NIMA's DNCPS program in great detail, and, given the opportunity, many more issues could be discussed. In particular, the elimination of unnecessary feature duplication between DNC libraries firmly deserves a mention - maintaining the effectiveness of any database is, after all, as much about getting rid of the dead wood as anything else.    However, we have managed to touch on several vital points. The future of mapping and charting, as we have seen, has now reached a turning point. It is no longer necessary to sit back and accept that disparate data sources can only be managed in an ad hoc, hope-for-the-best fashion. A unified, simultaneous, integrity-assured approach to geospatial management is now a reality.
    Neither is it any longer true that the application software should hold the key to your system's intelligence. Indeed, by making intelligent data available to multiple applications, you exploit lateral similarities rather than creating artificial differences.
    Separate applications need not replicate functionality, as this can be held in the database and shared among all applications. It is a fundamentally more economical route to mapping and charting success.
    Features or objects - call them what you will, they are the future of the geospatial disciplines. NIMA, like many other mapping agencies worldwide, is actively preparing for this future, while the users who prefer the reminiscences of outmoded technologies grow steadily fewer in number.

About the Author:
Simon Fletcher works in the Marketing Department at Laser-Scan Ltd, Cambridge, UK. He is published in numerous trade magazines and journals and writes in English, French, and German. He can be contacted on +44 1223 420414 (phone) +44 1223 420044 (fax) [email protected] (email)

Back