Posted on 11 January 2011 by Admin
When we devised the ‘Linguistic Geographies’ project in early 2009 we had in mind using an existing digital scan of the Gough Map undertaken in 2005, which was used in digitizing the map’s features as part of our earlier GIS-based ‘Mapping the Realm’ project. Fortuitously, early in 2010 the Bodleian Library undertook a new scan of the manuscript, as part of a conservation exercise, giving us a much better scan with improved resolution. It made sense therefore for ‘Linguistic Geographies’ to use this new map-scan, rather than the first, so the zoomable image makes use of this. However, what we found is that the map image of the new scan differs ever so slightly from that of the old. Why is this important? Well, the digitized map-layers developed as part of the ‘Mapping the Realm’ project now no longer ‘fit’ over the features as they are shown by the new scan, as can be seen on the right. Although the difference between the two scans is very small, they are nevertheless noticeable, for example in the way the rivers ‘vectors’ (seen here as purple lines) and ‘routes’ (in blue) do not match closely the rivers and routes depicted on the scanned map. So to make use of the ‘vector’ data from the ‘Mapping the Realm’ digitization work requires some editing, something undertaken within the GIS software (ArcGIS) used by the Queen’s University team. To do this, the new map-image is first ‘georectified’, that is fitted as closely as possible to the GIS data derived from the earlier scan (using a second order polynomial transform within the software), a process based upon selecting a group of key ‘control points’ distributed across the map as a whole. The software then uses these points to ‘stretch’ the scanned map to make the new image sit neatly on top of the first one. This automated process only takes us so far, as shown in the image above. To get the digitised map features to fit more closely requires manual editing, too, which is essential but rather repetitive work. The gain by having the new scan of the map is, in the shorter term, somewhat tempered by having to undertake this additional and unanticipated digitization work, therefore. The ultimate advantage of all this, however, will be to have not just a high-quality map image available online but also the capability to allow users the option of looking at the map’s cartographic features separately from each other (and also independently of the map image), so making it possible to examine and visualise, for example, the patterns of rivers or routes that are shown by the Gough map.