A place in history: a guide to using GIS in historical research

CHAPTER 3: ACQUIRING SPATIAL DATA

3.3 Digitising maps to produce vector data

While scanning produces a copy of the source map, digitising extracts certain features from the source and creates point, line, and/or polygon layers from them. From an early series of Ordnance Survey (OS) inch-to-the-mile maps, a user might simply want to extract a layer of points representing the locations of churches. Other layers, such as the road and rail networks, administrative boundaries, and so on may also be extracted. However, it is not possible to create a direct copy of the source map in the way that scanning does.

Points are digitised by clicking on the features that are required. Lines are digitised by tracing along the lines, clicking at points where there is a significant change in the line's direction. Polygons are created by creating a topological structure on top of line data as described in the previous chapter. Digitising can be done either directly from the paper map, using a digitising tablet or table, or by first scanning the map and digitising on screen, known as head-up digitising. Digitising tablets and tables (tables are usually larger) consist of a flat surface on which the map is firmly stuck down. The tablet or table's surface has a fine mesh of copper wires underneath it. There is also a puck, a hand-held device with a fine cross-hair and one or more buttons. To capture a point the cross-hairs are placed over the feature of interest and a button is pressed. The tablet or table is able to calculate the exact location of the cross-hairs from this and the coordinate is passed to the GIS software. With head-up digitising, a cursor is placed over the feature on the screen using the computer's mouse and a button is clicked to determine its exact location. Head-up digitising has the advantage that it creates a scanned copy of the source in addition to the vector data extracted from it. This both preserves a copy of the source, and can be used as a backcloth that provides context for the extracted vector features.

Digitising accuracy is extremely important especially where topology is to be created. A node will often have to be digitised two or more times to represent the end point of one line segment and the start point of another. Most GIS software provides tolerances that mean that if two nodes are within a set distance they will be 'snapped' together to form a single node. If the tolerance is set too high inappropriate features will be snapped. If it is too low then gaps will appear known as dangling nodes where polygons will not close properly. This will lead to corrupted topology.