TransDEM Forum

I have found that manual terrain adjustment for the area I am modeling would be never-ending as it is a mountainous area! I thought I had a solution to creating semi-accurate 3d vectors for track spines, but come to find out that I don't know how to calculate the missing elevations and thought that maybe someone could point me in the right direction.

Here is my process thus far:

I have literally traced a fair amount of trackage in my area on google earth and saved those paths as KMLs. I was then going to physically go to key locations along the routes (IE grade crossings, etc...) and get an accurate altitude reading with my GPS/altimeter and add these readings into the KML at the proper locations.

From this point I am unsure how to calculate the missing altitudes for points along the gradient created between the measured points. I am sure that a CAD application can calculate this information, but I am not too familiar with CAD software and I really don't have the money to afford any that I have found. Does anyone have any possible solutions to this problem or can recommend a better approach?

I can post some example KML data once I get home, if anyone would like to see what I am attempting!

Thanks,

Jeremy

If you are using an outdoor GPS unit with an barometric altimeter, you may be able to record the track in 3D pretty accurately, good calibration provided. Your GPS unit may be able to export the track-logs in .gpx format. (Without advertising a specific brand, the Garmin units are quite good at both 3D logging and exporting as .gpx.)

TransDEM can read .gpx and you can export to something very basic like .xyz. .xyz can be opened by a spreadsheet program as a variant of .csv and you can apply a linear of weighted interpolation to your data with the means of spreadsheet math functions. Then re-export xyz again (as csv with a profile that places spaces for commas) and open in TransDEM.

G'day jnpenley,

As geophil is the man who does know, I would take his suggestions seriously. However, I spot a small flaw in your methodology, as good as it is, which assumes (albeit unsaid) that the gradient of the track is constant between any two given level crossings. This WILL most definitely NOT be the case...

...accurate details of the relative movement of the gradients of any railway line should be available from the respective Gradient Diagrams published by the railway company concerned...

Jerker {:)}

I do have a GPS with a barometric altimeter and I love it. On a day with fair weather it is actually quite accurate. Bad weather thows it off a bit, but I won't be out in any thunderstorms gathering data!

I would love to actually walk the routes and get a very accurate track from my GPS, but I really don't want to risk being cited for trespassing on the Railroad's right of way. The fines are quite hefty and enforced. Maybe some late night excursions could yield results!

I have thought about the problems imposed by using a constant gradient between set points. Even with points only taken at just grade crossings there would only be several sections of track that would leave me with a gap larger than about a mile (1.6km). I could easily gather additional points along these sections that are close to roads/trails and easy to access. I was actually shooting for about a 1/2 mile between measured points. It wouldn't be a perfect track, but should be accurate enough to create a realistic representation, wouldn't it? I just figured that any areas that didn't look right I could fix manually or go grab a couple of extra points in the effected areas and reprocess the data. I am just experimenting at this point, so nothing in this process is set in stone!

Of the two railroads that operate and maintain trackage in this area, one bluntly stated that they "Do no provide trackage information to the public" and I have been unable to get a response from the other. The only track maps offered by either railroad are small gif images that are not even to scale or accurate. I am actually very disappointed about this!

I am also trying to get a response from someone in the local surveying community as well for ideas, but I am yet to get a response from anyone.

Thanks,

Jeremy

G'day jnpenley (Jeremy, if I may be so bold),

It is a shame that the railroads concerned are less than 'forthcoming' and 'uncooperative' in their attitudes. One possible way around this would be (if at all possible) to befriend one of their employees (preferably, in this case, a driver/engineer) and to then get the information you want from an 'inside source'). That notwithstanding, given that you are generally looking at mostly 1/2 mile (800 meters, let's say) between elevation points, then things aren't going to be quite as 'bad' as I imagined. Eight Hundred meters is roughly the length/width of a Trainz baseboard and given a maximum gradient of 1 in 50 (2%) (assumed), this would mean a rise/fall of only about 50 feet (16 meters) over that distance. There really wouldn't be too many intermediate 'severe' changes over this distance, in a real situation. I suspect you would easily have a very close representation of that which you seek, with the 'measuring premise' you are using. Incidentally, where it is practicable, stations/depots are another excellent location from which to gather GPS data...

Jerker {:)}

Jerker,
I want to thank you very much for your input. Those figures are about the same as the ones I calculated as well and in the mountainous terrain I figured the inaccuracy would be acceptable.

I did get a response from a local surveyor about calculating the missing elevations and he seemed to be partly stumped, but said that he would look into it for me. So I sent him an example of a route. So maybe he can find a solution and I can test this method and see how it turns out.

Thanks,

Jeremy

For the German "Zusi" simulator I created a tool named "Absteckrechner" for track laying that produces accurate geometry, adhering to the standards of German Federal Railways (DB), presumably not much different to other railways' standards.

Disadvantage with the tool is that it has a cumbersome user interface, numerical only, requiring quite a bit of expertise. And documentation is available in German only. Therefore I guess it's impracticable. Theoretically, however, it would do the trick.

I have attempted to figure out a solution using a weighted interpolation as recommended by GeoPhil but failed miserably. I don't really even know where to start. Some of the calculated figures would have put me on the Moon!

A CAD application is probably going to be my only option and I don't even know where to start there either.

GeoPhil,
The application you mentioned in the previous post looks impressive. As you mentioned though, I don't know any german and from the looks of the description on the page the application edits all values that are not within set standards. I am looking to just calculate Altitude, but it is very similar to what I am looking for.

I will keep digging and will post if I can come up with something useful.

Thanks,

G'day jnpenley,

Jeremy, I have just recalled that Trainz user and creator (and erstwhile DEM expert), RPearson (otherwise known to the community as "Rob"), was doing the sort of thing you are attempting here, quite some time ago using AutoCAD. I would surmise that he would be the person most likely to be able to shed some light on the subject for you. As far as I know (but don't quote me on this), he is still 'active', so perhaps you could send him a PM through the Trainz forum and see if he is as cooperative now as he used to be...

Jerker {:)}

I want to thank everyone for the help and input. You have certainly pointed me in the right direction as I have finally found a solution.

After testing a demo of every Mapping and CAD application that I could find that will handle 3D Geo data, I have finally found one that will easily interpolate elevations into my routes, Global Mapper 12 (I'm not sure if previous versions also have this functionality!). It offers the following two options for completing elevation data while viewing the route information:

1. Evenly spread elevations between first and last elevations.
2. Interpolate elevations to replace 0 values.

Option 2 works perfectly on the data that I plan to use.

It also has a function to complete elevations from loaded DEMs, this may be useful if you are using 1/9 Arc DEMS for your terrain, but using 1/9 arc in TransDEM seems to handle this pretty well already.

The only downside to Global Mapper is the cost, at 349 US, it's a steep price for me to only gain this one function. So I may find myself being rushed to collect my route data before the Time Limited Trial expires (I don't know how long that is yet, I haven't gotten the Trial key yet!). It may be worth the cost if I can find another use for it though.

I hope this information helps someone else, as this has been a headache for me. Now I have to uninstall some 30 applications off my PC.

Again, thanks for all the help