LinesMan provides trench maps which are 'geo-referenced' for the first time. This means that when connected to a GPS on either PC or on one of the many PDA devices (Personal Digital Assistant -not included but available from us separately) the system will display the user's real time position over the map. This functions with trench maps, or with IGN maps or aerial photos for normal navigation.
Here is a photo of the system in use, note position marker (red dot/circle), and track (red line). You load the maps you want before you go, by a simple 'send to PDA' function.
Here you can see GPS tracks over the Trench Map, individual trenches can be located accurately
Q: How easy is it?
A: It is very easy indeed. However, it is entirely possible that many of our interested customers do not own a GPS, nor do they have a PDA.
So here is a brief description of the kit required.
Definitions:
GPS (Global Positioning System). This is an American military system of satellites in geostationary orbit round the planet which emit very accurate time signals by radio.
In widespread use, both in the military and civilian, is a GPS receiver that reads the time signals and works out where it is on the planet to within 5 meters, by comparing very tiny differences between the received signals.
The GPS outputs a continous datastream containing position information.
Plug this data stream into a PC, or a PDA, and the data is read, and the computer immediately knows where it is.
A PDA is just a mini PC, with more simplified software which makes it portable.
Many PC's, PDA,s and GPS devices are on the market, but in essence they are all very similar.
You may or may not own some, all, or none of this kit already.
The geo-referencing of these maps has been a long project. It is not simply a case of comparing a modern map with the old, as they were produced in very different map projections. A map projection is the representation of a curved piece of the surface of the earth on a flat piece of paper, and there are many mathematical solutions for doing it. The problems, and solutions of 90 years ago were also very different.
The Belgian maps were drawn using the Belgian Bonne projection, the primary triangulation of Belgium being done between 1851 and 1873, and the work completed in 1888. The Belgian triangulation was connected to France and was tied to the principal triangulations of Great Britain, by cross channel measures as a result of Anglo-French and Belgian co-operation in 1861-2.
The spheroid of reference for lattitudes, longitudes and azimuths was the old Delambre system devised in 1820 on the meridains of Paris and Peru, and became known as the spheroid of the Carte de France.
The Belgians adopted the Bonne projection, and used the same spheroid of reference, and used as the initial meridian the longitude of the Old observatory. The initial parrallel being 50 deg 24' of North lattitude.
The British Army inherited this system in the early part of the war. They knew at the time, the limitations of the Bonne projection, (for it preserved only bearing and not distance, causing the artillery many problems) but were unable to convert to a better system, due to limitations of resource. They adopted the French Lambert system after the Great War. (co-incidentally still used on the modern IGN). They therefore extrapolated the Belgian system of maps south across France, as an expediant. Its limitation had to be accepted and worked with.
They
then set about marking them up with the trench lines that were being dug
continuously.
The problem then was to take all this information and come up with a solution to convert the legacy of British Army maps to a functioning GPS interface.
Some people have tried the known point comparison method, whith some success. But for the numbers of maps that we have been doing, this method is not practical, or accurate enough. It relies on cross referencing old points with new, 4 points per map, old and new makes 8 points, x 1200 maps = 9600 points . If just one church or crossroads has moved, or is recorded wrongly, the method is no good, and this is very likely, as can be shown graphically with moved junctions etc, even churches have often not quite been put back in the same place. Also it works only on localised areas, where the projection differences are negligable enough to be ignored, but the further away you go from the matched point, the greater the error, and errors over distances become significant.
The method employed has been to go to grass roots cartography, With the assistance of Jason Spiller, who is a consultant to the telecoms industry with specialist knowledge of cartography and mathematics, a mathematical transform between Bonne and WGS 84 has been devised. With the help of some detailed research into the mathematical base, he was able to produce a computer program, which enabled the conversion to be accurate on a repeatable basis. If one is to pass a system on to others as a satellite navigation system, one has to be confident in the mathematics, and confident enough that it will work from one end of the country to the other (especially if you don't live there to check it). It also needs to be quick and reliable. A mathematical transform was the only sure way to achieve this, and we believe this is the only computational system that has been used to date. This sets our product apart, as being totally unique.
We
have therefore put the hard work into the front end, and have produced
a co-ordinate transform system, that will geo reference into any software,
without the need for the system to recognise Bonne projections. We have
been testing it fully since its initial trial in 2001 in MapInfo software,
and found it to be accurate right across Belgium and France. Having converted
to Memory-Map, it has further been tested in the field by a battlefield
touring company, Kingshead Adventures http://www.kingsheadadventures.com on an almost continuous basis for the past year, with excellent results.
Here there are many factors.
Having now tested many maps in France it can safely be said that the 1918 maps are far more accurate than the 1915. The British Army refined them constantly from aerial survey. This partly accounts for the numerous editions.
It is also unquestionable that the British side of the line is better mapped than the German, as obviously it was possible to allow direct survey.
Don't forget that the maps in paper format, have not only survived a war, but also nearly a hundred years of storage, either flat, folded or hanging. Therefore some may have stretched, or shrunk.
Additionally the scanning process turns a map into a pixel image. Sometimes the map is creased. The transfer process may induce slight errors.
It must also be remembered that these maps were merely drawings, and drawings often contain errors.
Modern GPS's can also return some mad locations sometimes, especially if located near trees or buildings, or when travelling at speed.
Having said all that, to sum up, the maths at the time was good, and considering that GPS could never have been conceived when these maps were produced, I have been amazed at how accurately they interface with a WGS 84 Satellite Datum.
Accuracy on average varies between approximately 3 to 10m. The main apparent variation being the fact that many of the mapped roads were single track in 1916, but are now built as double width carriageway, so often an impression of inaccuracy can be given.
A small number of roads have moved slightly, much of the mapped areas were obliterated by shelling, subsequent development can potentially put roads where roads were not before. But in general, this is not significant. See the Troubleshooting section for a more graphical illustration of this.
© Great War Digital Ltd 2006 | Terms & Conditions | Privacy Policy