Converting Trench Map Positions to Google Maps

[WhiteStarLine]

Hi all,

I'm looking for some map enthusiasts to help me verify a technique that in time I can extend to other trench maps and make available to forum members free of charge. For some time, I've read pages in a unit war diary and wanted to see the location on Google Maps. I also wanted latitude / longitudes so that I could have waypoints for a moving map GPS.

I've finished the first map and I'm looking for someone to provide constructive criticism. My aim was, given any trench map co-ordinate on this map, locate it on Google Maps so that a user could be be within a few metres. For example, a known location such as Bray church or Bonfray Farm should put the user on one of the church walls or roughly on the farm survey point. The design goal was that the user should be able to use a familiar tool such as Excel and have a location such as 65DJ19b40.30 appear as a hyperlink to Google Maps (GM).

To do this, I downloaded the full 15 MB trench map for 65D, saving endless scrolling, resizing and retiling online. I found 12 survey points on the trench map that I could locate as latitude longitude co-ordinates on GM. I plotted each lat lon as Bonne, Normal Mercator and Lambert Conformal projections using Snyder's technique, ensuring the inverse x,y gave the same lat/lon. I also used the full trench map JPEG to come up with an x,y co-ordinate, inverted the y axis and rotated each point through the grid-meridian angle. I then took the trench map distance and bearing for these points and mapped these against 6 separate reference ellipsoids and a unit sphere with radii from various models using Vincenty direct and inverse. Not surprisingly, closest correlation came from the Normal Mercator Sphere, Delambre (Carte de France) datum for the forward and WGS-84 for the inverse. My anchor point then became the ruined moulin de Rolmont (49.997, 2.53171), as it is a 1918 survey point and the outer wall and centre appear clearly on Google Earth. This gave me a test case of a 15,460 metre forward bearing to Bonfray Farm and the back bearing got me onto the wall of the windmill after a 30 kilometre return journey. Seven of the others were similar accuracy and 3 churches were no longer in their 1918 position. The only puzzle was a Somme lock, which put me 100 metres downstream, so maybe the lock has been rebuilt in a new position.

I then took 14 unit war diary locations and placed them into Excel. All were features such as Heilly or Vaire Chateau and appeared on Google Earth at an acceptable location. So, anyone who understands the previous 2 paragraphs is welcome to the spreadsheet and I will make any modifications and move onto other trench maps. As I said previously, I am happy to share the work as it is already assisting me follow references to locations that the 11th Australian Infantry Brigade were fighting literally on this day on the Somme in 1918. The Excel spreadhseet also allows the user to either input a trench map reference, or locate a point of interest on the map and put in the x and y pixel. These are converted to x,y in yards using the latitude secant and converted to WGS-84 spherical latitude longitude values suitable for Google Maps. As the result is hyperlinked, the internet page, on a machine with two monitors, can be seen simultaneously.

If you are interested in helping, contact me. The spreadsheet has minimal error checking and expects the reference to be put in using the J19b40.30 style syntax. However, if using a JPEG, the x and y pixel location are all that is required.

The attached image shows the 11th Brigade's account of the downing of von Richtofen and their intelligence officer's co-ordinates. I superimposed the trench map over Google Maps and the green icon shows where my application thinks the machine crashed. This is exactly where the eye witnesses put it and approximately 120 metres from the roadside sign.

Finally, for others wanting to verify this, my blind alleys were:

* Reverse engineering trench grid x, y co-ordinates straight to a latitude and longitude using Snyder's formula was not fruitful and I experimented with several origins such as the Antwerp observatory, Brussels (Uccle) observatory, Brussels gazeteer, Charleroi and other guesses without success. I was close to giving up and going for an Affine transformation when I thought I'd systematically try every reference spheroid / ellipsoid.

* A single, accurate survey point is a better option than a triangle incentre, as this approach got me nowhere.

* I almost abandoned the entire project in frustration until I rotated each co-ordinate by the surveyor's grid angle and suddenly, systematic correlation started to occur.

[RobertBr]

I have an Excel Spreadsheet to convert Trench Map co-ordinates to Lat/lon. I created it prior to my 2008 trip to Arras.

It did a pretty good job on the MAROEIUL Map (51B NW). It relies on having a known reference point, ie Lat/Lon for a known Trench Map co-ordinate, it then works out a range & bearing from that point to the co-ords of interest and hence a new Lat/Lon. So it was a matter of locating a recognisable point on a Trench map in Google Maps and hence the Lat/Long and co-ords.

The Range/bearing and Lat/Lon formula are aerospace based and well proven.

Primary errors are in the accuracy of the recognisable points, the projection of the Trench Maps and distance from the reference point.

My main use was to help me find points of interest when driving around the area. I only had a basic car Sat Nav from which I could read off Lat/Lon at that time. A pre-prepared French Blue Series map and a set of Lat/Lons allowed me to trace my Grandfathers Battery positions in the area.

I have not used it for a while but you could use it to check your results; we could argue about any differences later.

Bob

[Black]

Geeeeeeeeeeez Bill!!

I AM impressed! There was I feeling rather smug that as a standard-issue grunt-person that I could use a topo map and one of those prismatic compass thingies to navigate from Point A to point B via Points C, D and E etc (by day and by night in all seasons, weaher and terrain); perhaps even identify a target on the ground, call for fire from my supporting artillery chappies and adjust it; then possibly even crank up a defensive fire plan, but you take the cake!

There's words there that I ain't never seen before. AND you do that addinup&gazinta stuff too! Silly me - I thought a Snyder was just a muzzle-loading artillery carbine variant; but no, you just wrecked my next presentation to the Flat Earth Society. Hoo-eeee! Frightening stuff, Bill. We must talk.

On a more fortuitous note, I logged in this evening with the specific aim of starting a new post to ask if one of our more clever GWF colleagues could match a trench map to Google Earth for me, and there you are... I have client who wants to know about a 1916 trench raid conducted by the 6th Brigade AIF in the Fleurbaix area. I'll need to scuttle away and find the grid references again, but I recall the relevant map is 36 NW 1:20,000 Edn 4. I don't have Linesman and I could not find it in any of the other on-line catalogues that I usually check.

PM me, please and we'll see what can be done.

Regards,

Black

[Smithmaps]

Sometimes I despair.

The hours I have spent, pondering mathematics, geo referencing by hand over 1,700 trench maps.

The money outlay to bring a Great War Moving map display product to market. The time spent on research, Deals and contracts with the IWM, testing the maps the full length of France, building web sites, on line payments and advertising, and yet a thread like this appears, which is a discussion between interested people, who seem to know nothing about it, or me!

LinesMan will do this conversion for every map on the Western Front, and put the whole thing in 3D if you want. You can walk the fields with all those maps scrolling before your very eyes.

Every forum member who has the software, can do conversions for other forum members for free already, on pretty much any map printed by the British Army!

The co ordinates it gives can be viewed on Modern maps within the system, or can be transferred to Google maps.

I do not aim to dispel the cleverness or effort on this thread, but

I'm going to kill myself.

Guy

www.greatwardigital.com .....LinesMan

[Havrincourt]

Now now Guy, dont make a mess on the carpet

I have linesman and it does what it says on the tin....well box thing

[Chris_Baker]

That is a great post, Guy ... are you still there ... Guy? Guy? Oh dear.

[Ken Lees]

I have Linesman, have just loaded it up onto my new laptop and despite loading everything onto the hard drive (or so I thought) I can't find anything. For example, looking at the IGN1:250k map and clicking on 'show maps at cursor' I get no responses. I share Guy's frustration, but for totally different reasons!

[Simon Mills]

Unfortunately I don't have Linesman as I haven't yet reached the stage where I need to get seriously involved in that line of research. That said, thanks to a few generous people on this forum and a friend with a bit of knowledge about how Photoshop works, I have been able to put together a useful composite so that I can more or less pinpoint where one of my great great uncles was killed in June 1917.

If I can get a similar map for the trenches around St. Eloi in February 1915 then I hope to do the same, but while something like Linesman does sound fascinating, unfortunately I just wouldn't use it enough to justify the cost.

[towisuk]

I have Linesman, have just loaded it up onto my new laptop and despite loading everything onto the hard drive (or so I thought) I can't find anything. For example, looking at the IGN1:250k map and clicking on 'show maps at cursor' I get no responses. I share Guy's frustration, but for totally different reasons!

Ken,

Have you Clicked on "map" top left of screen, then chosen "Refresh Map List" then told it where to find the maps?....

regards

Tom

[WhiteStarLine]

Guy, thanks for the effort you have put into Linesman and the readiness of Linesman owners to assist forum readers with georeferenced locations. What I sought to to do is to be able to translate a trench map location into a Google earth position, so that I could follow my grandfather's journey. To help me locate objects in the night sky in my public telescope demonstrations, I had previously spent time implementing US naval observatory algorithms for object selection and telescope alignment and I wanted to find out whether the the same general techniques were applicable to the transformation from one set of geographic coordinates into another. The methods in my first post (distance and bearing) were not sustainable over more than one map sheet, but reverse-engineering x and y offsets into Bonne, then WGS84, has paid dividends and across 20 map sheets (12,800 square kilometres) I can generally locate a position to what I call 'churchyard' accuracy - ie I can position a map reference in northern France or Belgium to be on or around that building or at least in the church ground. While this is not intended to compete with Linesman's accuracy, it is generally within the margin of error of the original trench map reference.

I will post some images shortly and a description of how, after generating 94,500 latitude / longitude co-ordinates (one for each sheet, grid letter, grid number and sub grid) I have been able to generate co-ordinates that enable me to take a grid letter (eg sheet 62D grid J) and import the image into Google Earth so that the boundaries align as well as any manual import I have seen.

Guy, this will never replace Linesman and I strongly recommend Linesman to all serious researchers. However, already I have been able to help some forum readers travelling from Australia to France who just want a quick guide to where a relative served. I am proud to have been able to do this and happy to share the techniques free of charge to the many Forum members who have helped me in my own research.

I hope Linesman continues to be successful and I hope it has repaid your efforts to-date.

Cheers,

Bill

Canberra, Australia

[Ken Lees]

Ken,

Have you Clicked on "map" top left of screen, then chosen "Refresh Map List" then told it where to find the maps?....

regards

Tom

Hi Tom,

Yes, I tried that, but there are still no maps

[keithfazzani]

Sometimes I despair.

The hours I have spent, pondering mathematics, geo referencing by hand over 1,700 trench maps.

The money outlay to bring a Great War Moving map display product to market. The time spent on research, Deals and contracts with the IWM, testing the maps the full length of France, building web sites, on line payments and advertising, and yet a thread like this appears, which is a discussion between interested people, who seem to know nothing about it, or me!

LinesMan will do this conversion for every map on the Western Front, and put the whole thing in 3D if you want. You can walk the fields with all those maps scrolling before your very eyes.

Every forum member who has the software, can do conversions for other forum members for free already, on pretty much any map printed by the British Army!

The co ordinates it gives can be viewed on Modern maps within the system, or can be transferred to Google maps.

I do not aim to dispel the cleverness or effort on this thread, but

I'm going to kill myself.

Guy

www.greatwardigital.com .....LinesMan

Just seen this Guy, please, please don't do it. I am a techno idiot I can't even understand what all the earlier posts are on about. But I do know that Linesman has added enormously to my, and those with me, experience of visiting the Western Front and also those long winter evenings with war diaries. Keep up the good work!

[Gully Ravine]

I have Linesman and have nothing but praise for it and the effort put in by its founders / developers. However, there are battlefields of great interest to others outside the western front and (for understandable commercial reasons) there is no equivalent Linesman offering. Not a criticism, but a fact. Hence, we are left to invent our own solutions and I feel these are to be encouraged. Keith

[6th Shropshires]

I can not afford Linesman but hope to get it one day, until that day comes I'll have to carry on doing it myself, which I do love doing.

[WhiteStarLine]

Three days before leaving France, I ironed out the last of some strange and annoying errors by changing the model from a spheroid to an ellipse. I was then able to use it to find trench map locations in Flanders and France. I used Google Earth, some overlaid trench maps and a laptop with GPS and we were able to put in a trench map reference and navigate to it on the ground. Here is the area I cover, 20 maps of 32 x 20 km each. The different colours represent whether the trench map reference is from a 1:10,000 or 20,000 or 40,000 map:

[Gully Ravine]

How accurate were your overlays / calculations in practice? Reason I ask is that I have been testing my own maps (GE screen grabs of my local area for test purposes) and have not been as accurate as I would expect once these have been uploaded to my satnav. Too far out to be of any use with trench maps. Keith

[WhiteStarLine]

Keith, I am writing up my techniques and associated accuracy and will have that posted within a week. In summary, over the last two weeks I followed the 3rd AIF Division using Google Earth, a GPS and my calculated overlays of trench maps in Belgium (Messines, Passchendale), down to Armentieres on the Lys and onto the River Somme, west to Bray, Peronne and Bellicourt. I also traced D company, 23rd Battalion AIF at Pozieres where my wife's great uncle still lies. This involved 6 trench maps. For detailed work, such as following a trench raid or battalion attack, I used the lat / lon quadrants I calculate automatically and postioned the letter extract onto Google Earth without alteration (eg map 62D letter J).

The GPS in the car got us down the motorway to the right town and then my laptop with GPS and overlaid trench map on Google Earth took over and found us every location we wanted. I also carried a datalogging GPS to record the route we actually took in case I needed to make adjustments to the algorithm suite on my return. We also had my wife's iPad with GPS and a spare.

I strive for 'churchyard accuracy' which means I aim to get the Google Earth pushpin up against the church building in 75% of test cases, when ideally it should be positioned on the church spire with an uncertainty of 21 square metres. Sometimes I am on the roof and sometimes I am in the grounds itself. However, the technique works, is based on a correct theoretical foundation and has been field tested and logged. Here is one example, from Sailly-le-Sec. The green pushpin was plotted while in Australia and reflects the predicted value of the trench map reference to a latitude and longitude plotted in Google Earth. [A minor mistake is that I realised I had actually used a 1:20,000 map]. The red line shows my tracks as I walked around the church and right up to the door at the base of the spire:

Note that my prediction shows me in the middle of the church, which is as good as you can strive for given you are using a scanned image of a 100 year old paper map as your starting point. Ironically, a local resident told me that the spire was reversed when the church was rebuilt in the 1920s.

Finally, here is the overlaid image, with the Google Earth imagery faintly visible below:

The actual church location and trench map location correlate, as can be seen if you run the slider back and forth in GE. This is what we used in France.

On my return to Australia, I added the functionality to reverse engineer, so that given a town or latitude / longitude I can calculate the relevant trench map.

I am happy to make available my findings for either independent verification or to assist others wanting to follow the same route, as it was simple, accurate and free. A later post will summarise this.

Cheers, Bill

[Guest]

Maybe Guy should offer you a job, before you produce a cheaper alternative to Linesman?

Linesman would be a nice Christmas present. It is expensive, but looks worth the money. It's just a bit out of reach for a lot of people. I see there are different offers available. from £99-£500+ Click

So I will probably go for a cheaper version at some point. Am studying it.

Mike

[WhiteStarLine]

Mike, I readily agree that Linesman offers so much more. When I was at La Boiselle a few weeks back we struck up a conversation with the landowner. Once she realised we were from Australia, she spent a lot of time chatting and part of this was telling me how impressed she was with the trench plotting done by Linesman.

The cost is out of my range so I came up with something that worked for me using free software running on a standard laptop and GPS. One practical example was plotting where I suspected the 11th AIF Brigade was actually located when MvR was shot down, literally above their heads. I walked down the hill carrying my laptop and literally stumbled down into the basin they described as being an old quarry.

[WhiteStarLine]

Here are some precision and accuracy results over a 96 x 120 km grid representing the major battles of Flanders and Northern France. For each map sheet of 32 x 20 km, the coloured pushpin represents the British Trench Map grid reference and its predicted location on Google Earth. The associated image shows how close the prediction was to the target, typically a church spire, cross road or bridge. The distance from the origin is shown, ranging from 22 km to 129 km. After some field testing, the approach appears sound enough to support any forum member wishing to conduct battlefield touring with Google Earth and overlaid trench maps; display WW1 locations on contemporary maps or reverse engineer town and village locations to find the associated trench map.

Sheet 27 Saint-Omer de Staple church 128,958 m. from origin.

Sheet 28 St Vedastus church, Menin 90,171 m.

Sheet 29 Kortrijk, bridge (should have been the centre) 84,072 m.

Sheet 36c Givenchy-lès-la-Bassée 105,547 m.

Sheet 36 Laventie church 106,194 m.

Sheet 37 Eglise Sainte-Marguerite, Tournai 67,432 m.

Sheet 44 Orchies church 73,460 m. from origin.

Sheet 45 Givry church 22,174 m.

Sheet 51a Noyelles-sur-Selle church 65,006 m.

Sheet 51b La Targette crossroads 105,522 m.

Sheet 51c Fosseux church 118,499 m.

Sheet 57a Marbaix 44,007 m.

Sheet 57b Audencourt church 66,936 m. from origin.

Sheet 57c Bourlon church 84,779 m.

Sheet 57d Aveluy Railway Bridge (should have been the centre) 118,187 m.

Sheet 62b Beaurevoir Mill (ruin) 81,032 m.

Sheet 62c Suzanne church 114,328 m.

Sheet 62d Framerville intersection 121,320 m.

[Graham Wright]

I am to visit France with my 2 months to follow in my grandfather’s footsteps to discover where he was killed on the 28th March 1918. I have map references but I am having difficulty in interpreting them within the relevant map 62D SE. A copy of the extract from the War Diary is provided below.

Any help would be appreciated.

Graham Wright 

Extract from war diary.rtf

[Guest]

I don't have much time just now. Hopefully this is of use to you. Input like this 62d.G.25.a.8.9. into The British Trench Map Coordinates Converter

 

Or use one of our forum pals site the excellent tMapper

 

Mike

 

 

 

 

[Howard]

Remember not to expect too much in terms of accuracy. If you geo-reference a map, pop it on your GPS and stand on a trench you may in fact be a few meters off despite great care. There are a number of reasons for this.

In no particular order:-

1. Trenches were often marked on a map by scaling from aerial photographs which themselves had the map grid drawn on them. Errors in placing that grid on the photo and in scaling and transferring the data to a litho plate will often give errors of a metre or two.

2. The aerial photos appear to show level ground but any slope, any deviation of the camera from the vertical, lens aberrations etc. will give a slightly distorted image. They knew a great deal about this (photogrammetry) problem and took steps to correct it but those steps cannot eliminate error.

3. If you look at the corners on many trench maps, will will see the printing registration errors, coloured trenches are over printed on stock maps, sometimes the error is 10 metres or more, sometimes spot on.

4. Domestic or hand held GPS is not perfect, sometimes errors of 5 metres can creep in.

5. Trenches "moved" after being re-dug after shelling.

6. Maps geo-referenced by using modern roads may be off as most roads and especially junctions have been widened. If that widening was to one side only, using a road centre as a georeference will give another error.

7. Maps geo-referenced from church steeples, bridges etc. can be off as many were destroyed in the war and may not have been rebuilt "exactly" as before or even in the same place.

8. Google maps geo-referencing is sometimes quite a way off. This is not uncommon, even the famous 6" OS maps of the UK are not always very accurate, they sometimes have displaced grids etc.

 

Mapping is hard, accuracy is elusive and it requires diligent attention to detail.

 

Of course it may all work and you may stand more or less where "it" happened, but don't be too surprised if you don't.

 

Howard

[Guest]

On 1/30/2018 at 14:33, Howard said:

Remember not to expect too much in terms of accuracy. If you geo-reference a map, pop it on your GPS and stand on a trench you may in fact be a few meters off despite great care. There are a number of reasons for this.

In no particular order:-

1. Trenches were often marked on a map by scaling from aerial photographs which themselves had the map grid drawn on them. Errors in placing that grid on the photo and in scaling and transferring the data to a litho plate will often give errors of a metre or two.

2. The aerial photos appear to show level ground but any slope, any deviation of the camera from the vertical, lens aberrations etc. will give a slightly distorted image. They knew a great deal about this (photogrammetry) problem and took steps to correct it but those steps cannot eliminate error.

3. If you look at the corners on many trench maps, will will see the printing registration errors, coloured trenches are over printed on stock maps, sometimes the error is 10 metres or more, sometimes spot on.

4. Domestic or hand held GPS is not perfect, sometimes errors of 5 metres can creep in.

5. Trenches "moved" after being re-dug after shelling.

6. Maps geo-referenced by using modern roads may be off as most roads and especially junctions have been widened. If that widening was to one side only, using a road centre as a georeference will give another error.

7. Maps geo-referenced from church steeples, bridges etc. can be off as many were destroyed in the war and may not have been rebuilt "exactly" as before or even in the same place.

8. Google maps geo-referencing is sometimes quite a way off. This

 

Hi I was pointed in this direction to help find where a relative of my mine died. The man was Walter Osborne Varian 2nd Lieutenant 2nd Munster Fusiliers. His place of death was Hamel at the Somme. This is the map where I found his coordinates: http://maps.nls.uk/view/101465329

 

His coordinates are 62d p9 b93. 

 

Any help on finding this on a modern map would be greatly appreciated!

[Guest]

If 62d.P.9.b.9.3.  is correct, I make it at the attached google earth location.   49.8998, 2.5692

 

Sorry had bother with that as a Kmz file, have attached image, if that's any use.

 

 

 

Mike

 

Edited 23 February , 2018 by Skipman