Most people are familiar with paper roadmaps. How are digital maps different and why are they so much more powerful? Paper roadmaps, and in fact all paper maps, have a major limitation in that the spatial "database" is the drawing on the paper. This imposes inherent limitations on the building and use of the map data. First, the amount of data you are able to capture on the map is severely limited by what can be clearly drawn and understood from the map. Second, the paper map is a static snapshot. It is not possible to change the map scale or easily update the map with new information. Finally, it is quite difficult to do quantitative spatial analysis with a paper map, such as calculating the fastest route from point A to point B or quickly looking up an address.
A digital map attempts to capture the underlying geographical phenomena and make it available for dynamic retrieval, spatial analysis, and representation by sophisticated software systems. There are two basic ways digital maps can represent what is present in the map and where it is. The first is based on the concept that the geographic world is composed of entities that can be positioned on the map by a geometric coordinate system and described by attributes and properties. This approach typically uses a vector data model, where entities are defined using points, lines, and polygons (see Figure 4.3 and Figure 4.4). The second is that specific attributes (e.g., elevation) vary continuously in the map as a mathematical function. Because it can be challenging to represent large geographical areas by a simple differentiable numerical function, it is common to divide the geographical space into discrete spatial units. The result is known as a tessellation, and can be composed of square cells if a raster model is used (see Figure 4.5).
The spatial analysis required in mobile location services (e.g., routing through a linked network topology; i.e., roads) is well served by the entity model. This is the most common map data model used by spatial analysis software vendors, and this is the data model focused on in the remainder of this book.
Building a digital map is a continual, highly labor-intensive, expensive process. The map vendors that produce highly detailed maps suitable for mobile location services (referred to as large-scale maps) typically update their map products between two and four times per year. The following steps provide a basic overview of the digital map creation and maintenance process.
Digital maps begin with raw data collected and aggregated from local vendors. Map data companies typically employ source acquisition specialists who continually screen multiple data sources for completeness and accuracy. Their function is to continually update relationships with local vendors to collect map source materials from public and private suppliers. These relationships include those with local government agencies, which provide updated materials such as aerial-rectified photos and differential GPS field surveys. This information allows the map data vendor to extend digital maps with information like new roads, postal codes, and address ranges.
After the core map database is created (or when an update is required), a group of specially trained professionals drive the roads to compare reality with the digital data, and to collect new features and attributes for the map. This information could be new turn restrictions, road geometry, and signage information. Additional information that might not be available in the core map database can also be collected. This could include one-way streets, exit signage, prohibited turns, tunnels, bridges, vehicle restrictions, and address ranges. This information is delivered to a production unit, who will make the necessary improvements to the core map database.
The source map data needs to be converted from its source format into various product formats optimized for specific applications. These product formats could include MultiNet GDF, Shapefile, MapInfo, MapBase, MapAccess, Spatial Data Engine (SDE), Oracle Spatial, KIWI, geocoding-specific formats, or lighter formats such as those for mobile location services.