This guest article by Nicolas Regnauld, Product Manager at 1Spatial takes a look at automated generalization and how this holy grail has come closer for national mapping and cartographic agencies. The process of creating lower-resolution and smaller scale products from a single high-resolution data source has been a very labor-intensive process. Automation means organizations can do away with interpretation methodologies and so-called “cartographers license”. Automation can provide continuous, up-to-date and resolution-specific digital landscape models (DLM) and style-specific, digital cartographic models (DCM).
Map generalization is the process used to reduce the scale of a map, to display a wider geographic extent on a similar size of paper. This involves exaggerating the more important features, like roads, so that they are readable at the smaller scale and removing unnecessary detail. As some features are exaggerated, space becomes scarce and cartographic conflicts (partial overlap between features) may occur. This requires careful shifting of features to obtain a readable map.
This was traditionally done manually in National Mapping and Cadastral Agencies (NMCA) by cartographers. This process is slow and expensive which is why many map series are updated once every few years. Many NMCAs have now invested in a new approach where data is captured once at very large scale and stored in a topographic database. The expectation is that it will be possible to derive a wide range of maps from it using highly automated processes. Automatic generalisation is often described as the Holy Grail for NMCAs, as it enables them to derive multiple products quickly and at low cost allowing them to quickly adapt their offering to the needs of the market.
Automating the generalisation process is very challenging though. Cartographers are able to use their brain to analyse every geographic local situation and find an aesthetically pleasing way to represent it on a map. Different cartographers will often produce different results for a same situation. An automated solution is less able to adapt the process for each situation, so the computer cannot match the cartographic quality that a cartographer can produce. However, a computer brings speed and consistency. Speed means that the product can be updated more often, and new products can also be created more easily. Consistency is particularly important when the product is a dataset that is fed to other systems that perform specific tasks (risk analysis, simulations, routing, etc.). These automatic processes often rely on the data meeting strict specifications, which are more easily enforced when the data are produced by an automated single process, rather than a number of cartographers performing manual tasks.
1Spatial has long been involved with the development of automated generalisation software. 1Spatial generalisation technology has been successfully used by NMCAs in Europe, to automate the generalisation processes in some of their map production systems. These solutions generalise geographic features according to a set of cartographic constraints, which represent the specifications of the target map. An optimisation engine is used to maximise the satisfaction of these constraints. These solutions also use a framework for authoring and running rules to detect and react to specific conditions in the data. All of this gives NMCAs access to a powerful, common object database with explicit topology management.
Generalisation enables NMCAs and other data publishers to automatically create multiple smaller-scale data products from large-scale data sources. It saves huge amounts of time and effort when producing data products and allows organisations to rapidly generate new products in response to market demands. Being able to automate the process means that organisations can improve the consistency of their data products and provides opportunities for them to increase the product ranges offered so that they can increase sales or attract new customers.
So we might have lost the art of the cartographer, but we are working towards the holy grail of providing users with always up-to-date information.