Automatic Generalization - a Case Study

swisstopo – The Customer

swisstopo is the competence center of the Swiss Confederation responsible for geographical reference data and all products derived from them. swisstopo’s National Map series enjoys international recognition.

The Federal Office of Topography (swisstopo), located in Wabern near Bern, is a modern production organisation with nearly 270 employees. swisstopo creates and maintains the basic geodetic, topographic and geological data for the whole of Switzerland, publishes the national map series at a variety of scales and keeps them up to date. The official cadastral survey, providing the measurement and mapping for the national land register, is a joint task of the Confederation and the Cantons and is controlled by a division of swisstopo, the Federal Directorate of Cadastral Surveying. The National Geological Survey is also a division of swisstopo. Over recent years, a National Spatial Data Infrastructure (NSDI) has been established in Switzerland to promote the interoperability of digital geographical data between the various suppliers and users. As the responsible body within the Confederation, swisstopo is heavily involved in this task, through its twin objectives of offering high quality spatial reference data and derived products and of increasing their use throughout the national economy. swisstopo was founded in 1838. Today swisstopo forms part of the armasuisse group within the Federal Department of Defence, Civil Protection and Sport (DDPS).

The Requirements

In recent years, swisstopo has recognized a need to update its production processes and enable a more rapid update cycle. In order to do this, the existing process that included a large amount of manual production work, needed to be re-considered and systems that allowed for higher levels of automation introduced. swisstopo identified a need to be able to derive several map scales from one data source. Also, they determined that the updating cycle, which is currently every six years could be shortened significantly.

After careful preparation, swisstopo set up a worldwide WTO tender process in 2005. Within the tender for a cartography production system, the project OPTINA-LK (Optimierung der Nachführung der Landeskarte) had the goal of optimizing the updating of national maps and included two separate blocks – one for automatic cartographic generalization and incremental updating, SysDab (system for automated data derivations) and one for manual editing, data maintanence and output. The strategic swisstopo OPTINA-LK project is designed to replace existing map production technology with a new and future-oriented GIS-based map production system for the creation, generalization, maintenance and updating of digital cartographic models.

Axes Systems' automated GIS cartography solution axpand best fulfilled the requirements of the SysDab WTO tender as listed in the requirements manual and offered the most economical solution. SysDab is a GIS-based system for automatic semantic, geometric and cartographic generalization of geo-data. SysDab will be used for extensive automatic derivation and incremental updating of digital cartographic models from digital landscape models or large scale digital cartographic landscape models.

Besides fulfilling the requirements and offering an economic solution to a complex problem, Axes Systems was selected to provide the automatic generalization for several reasons:

• Axes Systems’ wealth of experience in data-based digital cartography and the accompanying technology
• Axes Systems’ technology for and experience in automated generalization
• Axes Systems’ large number of customers already using axpand cartography and generalization for topographic map production
• Very high level of correctly generalized data to ensure minimum manual editing in the subsequent system
• One integrated data base for all derived cartographic data models
• Integrated cartographic symbolization and visualization
• True multiple representation and multiple models of geo-data MRDB
• MRDB which maintains all relationships between source and derived data objects (1:1, 1:n, m:n – relationship) as well as the influencing objects (horizontal relationship)
• MRDB which maintains relationships between old and updated objects over time (update relationship) to enable automated incremental updating and update relevance checking

axpand by Axes Systems

axpand, Axes Systems’ core product, is well-known in the market for its exceptional visualization and cartographic functionality which includes automated generalization. In 2004, Axes Systems began enhancing axpand’s existing automated generalization functionality, to include a more advanced level of generalization.The most recent automatic generalization in axpand rests on workflow technology which is a complex and integrated flow of combinations of generalization operators which use user-defined criteria (constraints). This functionality is based on a true multi-representation data base which allows for several representations of the same data to be stored in one data base. This, in turn, allows the system to store a history of the changes that were made to individual objects during the generalization process and utilize the relationships between the source and resulting data for automated incremental updating.

SysDab – System for automated data derivations (Ableitungen)

Axes Systems was awarded the opportunity work with swisstopo to refine its automatic generalization functionality to meet the generalization requirements necessary to derive data at various map scales while keeping the quality of the existing map series.

The source data originates from an external, high-resolution topographic landscape model (TLM), which is captured at an exactness of one meter for certain object types including buildings and streets. The data is imported to axpand’s multi-representation data base using Interlis, a Swiss standard exchange format for geo-data. The data is then generalized according to a detailed set of criteria which include several generalization processes that have traditionally been managed manually. swisstopo worked intensively with Axes Systems to develop this set of criteria which has taken on the character of a model in this area. The challenge here was to convert the cartographer’s knowledge about traditional manual generalization to logical rules that can be used by the system to produce consistent high quality generalization results. Once the constraints were defined, Axes Systems applied the logic to its automatic workflow technology and carried out an initial generalization of digital cartographic data at 1:25’000 scale (DKM25 derived from TLM) and a generalization of the landscape model Vector 200 (1:200’000) to digital cartographic data at 1:300’000 scale (DKM300). The DKM25 can be used to derive a 1:50,000 scale and the 1:50,000 scale to derive a 1:100,000 scale. The original system goal for SysDab was to achieve a 75% level of automatic generalization based on the defined criteria.Objects that conflict with the criteria are logged by the system and can be sequentially reviewed within the system. The high level of automation will, however, cut down the number of the cartographers’ manual editing hours to a minimum while maintaining the current high quality of swisstopo’s maps. swisstopo’s maps are internationally recognized for their unrivalled quality standard. The automatic generalization is configured to mirror the generalization that is produced manually. See examples of automatic generalization with various data sets using  axpand ng.