Digitizing Existing or Abandoned Lines

[landnrailroader]

As those of you that have downloaded my routes know, I focus my efforts on existing or abandoned rail lines. I have no problem with fictitious material and some great routes have been done that way, but it is not my area of interest. So one of the early issues one will face is just how to accurately digitize a route. Well, if it is an existing route, for example, my CSX J-Line, main, then it is easy, just put the dots in the center of the main track. Then jot down the coordinates of turnouts or other features on that line and enter them into the marker file.

I use Google Earth although in the early days of the hobby, there was a program written by an airline pilot called USAPhotomaps which was great if you were doing a route in the US. But that required a database that Microsoft did away with by 2006 or so. Since then, I have used Google Earth. The Pro version is still free and it offers the ability to import topo maps as an overlay. The "geotopo" maps are one of the options available when you use the USGS site(s) to obtain free topo maps. What you get when you down load one is a borderless map that is fitted to overlay an area in Google Earth. But it is not perfect so here is what I do. And this will apply mainly to abandoned lines.

Use the oldest imagery available in Google Earth which is usually monochrome, i.e. B&W and usually is from around 1997. Follow the rail line on the topo map overlay and digitize it for perhaps 20 miles. Then start a new file for the next 20 miles or so. I just do this for convenience. Now turnoff the topo map overlay (uncheck it in the left column). Now if you examine the B&W image, I find that the true line of track, usually appears as a lighter shade of line and it will be generally close and parallel to the line you have created. However it may be slightly off, and where those areas are, select each point and move that point to the new, lighter shade line. If you have a track chart, and let's say that this chart shows that there are no curves from mile 7 to mile 10, then another helpful trick is to place a measured line between two points on the digitized line that appear to be correct. Save this line (it will save in the temporary area of the left column). Then adjust the digitzed line to match this long tangent.

In the attached image, the colored line with dots is the original digitize from the topo map, but notice the brighter line roughly parallel to it. This is the rail line R/W on the 1997 aerail image. This is of the Columbia Branch of the NC&StL----L&N----CSX which existed in 1964 (when I rode the mixed train over it) but was abandoned in the 1980s. It is an area 3 miles or so railroad north of Fayetteville, TN. So I will select each dot and move it to the lighter gray line. Now the B&W image is not perfect but generally I find that it is closer to correct that later color images. You will see significant variations between aerials and if a line is obscured by trees or brush, then you just use an educated guess. And often I just digitize off the B&W image in the first place but sometimes farmer(?) has plowed undere the R/W which now crosses his pasture and does not show at all and then all you have is the topo..

DIGITIZE.jpg
Jerry

[MMCdude]

Thanks for posting this. When life finally slows down enough to allow me time to learn route building, my goal is to tackle the GM&O from Venice, IL (East St. Louis) to Jackson, TN (mostly abandoned now)...

[scfe_fan]

This is a pretty nifty idea. I'm working on an abandoned line that has been abandoned for so long that there is very little trace of it ever being there save for a bunch of old bridges. The one upside is that a highway was built over the old right of way and it was basically single tracked the whole 180-ish miles or so.

[CSXT Q-317]

I am working on digitizing the old DT&I from Ironton to Flat Rock in order to create more accurate track.

Is there any limit to how many of these marker files you can have in at one time?
And can you merge two of these into one file?

[landnrailroader]

M. Payne,

I have over 50 marker files as I work on the Chattanooga Sub. Now you can only have one in the route but using editor, you can copy/paste other marker files into the marker file in the route or delete portions no longer needed. So just for my ease of use, I limit my working marker file to around 4000 lines. For example there are 4 marker files for the Nashville to Chattanooga main, and I had all 4 together when I laid that track. Then I deleted most of those when I added 2 Sparta branch files. Now that all track is done, I will be digitizing the main roads and those that cross the track. When I deliver an incomplete route, I include the marker files and the files that can be loaded into Google Earth if necessary. I use Context editor which can be found on the "net" but I expect others would work too - just to be one that can handle unicode text files.

Jerry

[landnrailroader]

Collegues,
Here is another example of digitiziing. In this case it is the abandoned portion of the former Tennessee Central from Monterrey, TN to Crossville and beyond to Crab Orchard where Ljoist has a large quarry and operates their own train from that point to connection with the NS at Rockwood. There are 5 branches off the main line. One to Old Hickory, outside Nashville is still very active and serves many customers. Another that goes from Carthage Jct. to Carthage, I think is still active, at least some. Then there is a long branch to Wilder (Coal Mine). Isoline (agriculture & Coal), and Dripping Spring (Coal). These last 3 have been gone since before 1950. However they are shown on topo maps which we know are not all that accurate relative to aerial images but it is all we have. So here is a situation where the blue line is the digitized line from the map, and the red line is a highway digitized from the aerials, both of which are pretty close. But note that in the B&W image, the light gray line below the blue lilne is quite visible but in the color image nothing is to be seen in the area of the trees. CROSSX01.jpg CROSSX02.jpg
So I will adjust the digitized blue line to better fit the B&W image and run with it-----.
Jerry

[eric]

What I've found on abandoned lines is tree lines usually give away the ROW boundaries pretty well.

[oper190]

So I will adjust the digitized blue line to better fit the B&W image and run with it-----.

Jerry just an FYI -I was able to superimpose the B&W over the digital image. As seen, when place the B&W over the digital image, as a layer so I could fade it in to see though, it is obvious the scaling is a little off between them. I aligned the images with the buildings in the upper left, and there both your blue line are almost an exact match. Note that the road and the mapping location markers are off. if I aligned on the road then the buildings shift upward, but then your blue line on the digital mage aligns with the gray right of way on the photo by the road. Therefore I'd say the digital alignment is correct.

CROSSX02a.jpg

[landnrailroader]

That looks pretty good but it is still off by several feet in the curve. I decided to go ahead and do the whole route from Harriman to Nashville since it was my first experience at riding a locomotive - back in 1955 or so, when I was 16 and a friend's grandfather was the engineer. I will be doing per the period just prior to 1969, but it will include all existing customer connections (sidings). Now the line from Nashville to Monterrey is a Corman owned shortline, no track between Monterrey and Crab Orchard, and then an industrial road from there to Emory Gap, but essentially ending at the NS interchange in Rockwood. Since we never know when our PC will "burp", I have also laid out the L&N lines from Nashville south to Montgomery, AL which was the line of the S&NA road. The South & North Alabama RR had two parts, Nashville to Birminghma, and Birmingham to Montgomery. Given my age of 87 in other week, I may never actually work on that route, but my son-in-law has the information he needs to upload any of my unfinished routes and then Eric could retain them, or dump them depending on perhaps how many downloads there are in some period of time.

On another issue, how do we know the details of a "real route". Used to be, Railfandepot was a popular site where one could download .pdf files and such of just about anything you wanted, and now there are still dozens of line out there on CDs or DVDs. Railfandepot is now handled on Ebay, so all you need to do is go to Ebay and search for what you want. I recently got the BNSF Oregon lines because one of my "to be done, maybe" routes, already digitized, is the SP&S from Wishram to Bend, OR. This CD cost a whole $12+ amd shipping was free. Then another source is the historical society of the line of interest. The L&N Historical Society has a web page and in the "store" section there are a number of books, CDs etc. for purchase. Members get a small discount. So, a few days ago, I ordered 2 CDs. One contains this old Tennessee Central line as far as Crossville, and the other has the lines between Nashville & Montgomery. I believe someone has done the M&M Sub. south of Montgomery, years ago, and also at least part of the NO&M between Mobile and New Orleans. These CDs are $25 each plus shipping, total of $55 and they accept Paypal as well as your magic card(s). The much newer BNSF track chart is interesting in that I can tell at a glance that it was done with Microstation (CAD), but of greater interest is that the elevation above sea level is show at each milepost and for the most part, gradients are shown between mileposts and do not have the infinite variations found on older charts - which you probably can't duplicate in ORTS anyway. The Southern Rwy. historical group shares space in Chattanooga, TN (at the TVRM) with the L&N group and both groups jointly employ one person to handle these orders. I've been a L&NHS member for nearly 30 years and occasionally order things from the Southern group - I was a SR employee from 6/65 to 8/78 when I moved to the FRA.

Track charts also give you the degree of curve and since I use M-tracks, this is helpful. In some cases, signals & type of turnouts are also shown.

This Nashville to Birmingham route includes the old "Peavine" from Nashville-Columbia-North Athens which was the original L&N mainline (S&NA) and also the revised new line from around 1905, Nashville-Lewisburg-North Athens. It will also include a short piece of the Columbia to Mt. Pleasant line and someone could expand that to extend to the Sheffield, Al area. This was also L&N, but now from a short piece north of Columbia to Sheffield, and Columbia to Pulaski is now the Tennessee Southern RR, a Patriot Rail property. So one could do it as old L&N, or new TSRR.

Jerry Sullivan

[tarrant75]

Jerry, I absolutely applaud your efforts, and I can't thank you enough, not only for the work you do, but for the teaching of how to do things.

I've long wanted to do the PA cement area in around the Lehigh Valley, as well as the old lines as they were in the area before CR gutted it in the late 70's/early 80's (such as it is in the "Railfanning with the Bednar's" DVD series), but finding track plans is heard enough, let alone figuring out where the lines were due to a lack of imagery from that era, and so much changed from then to the start of the aerial imagery being recorded

[landnrailroader]

Chris,

I can imagine that would be difficult. I am not familiar with the Lehigh area although a long time friend of my late wife and I lived near Emmaus, PA and passed away a couple months ago. Then my wife (now of 32 years) and I plan to take a Kelly Tour in the fall that will involve riding a tourist line in the Lehigh area as well as the Strasburg.

Even when you can find a line on some source you can digitize from, (my favorite is Google Earth) there are issues. For example, GE was apparently not intended for precision digitizing like we want to have but it is free so ? There is historic imagery going as far back as 1997 (B&W usually) but the various data sets do not precisely line up with each other, and may be anything from a few inches to 50 feet in error, and then you are digitizing a track line and it is great for miles, and then there is a sharp offset of 10-15 feet? Well in those cases, I just average it out by eye realizing that if it is in tangent track, I just use a piece to intially set direction that will bridge this anomaly, or if it is in a curve, I will lay the curve to match the end point tangents. Lastly in laying the track, for example one stretch I am currently digitizing is tanget for more than 5 miles, so? What I will do when I lay the track is I will complete the previous curve using itty bitty pieces. For example, we have fractional pieces of 1/4 & 1/2 degree in the X-tracks, and 0.1 degree in M-tracks, so I will adjust the starting curve and use 500M tangent pieces to align to the tanget and for this one I may use 5-10 500M pieces until I am satisfied with the alignment. Then I will lay the tangent that I will save with shorter pieces where the track crosses a boundary. We found out years ago that it was not good to use much longer than a 100M piece across a boundary, and I guess that rule is still in effect.

When I am digitizing from Google Earth, I first look for two or three datasets that line up correctly, or at least provide enough accuracy that I can put my data points consistently within the gauge. I have a fairly large monitor, 15" high by maybe 25" in width. It is a ASUS product and it is the largest that I can fit into my available space. Bought it about a year ago in a STAPLES sale for about $200. So I zoom in until at least in most data sets I can see the individual rails. And I zoom in to the point that these are 1/16" or a little further apart and I place my data points as close to the center line as these 87 year old precision instruments (eyes) will allow. You can always adjust the last point you placed but if you go back further than that, and then place new points, the result is a cob web of lines on the screen, though the points work okay. However, after I have digitized 50 miles or so, that being my more or less maximum for a single set of data. I go back and adjust any that are way off center before I save the data. Save it someplace as a .kml file and then there is a utility in the library to convert .kml files to .mkr files.

Another issue is merging existing routes. I have not had success doing that but I have used existing routes to start new ones and that is easy. Just create the new route, preferably with MSTS/RGE if you have (if you don't, I have no idea what to do) it. Then copy the world files from your existing route into the new route. Then, manually edit (I use Context Editor) the world files and carefully remove any dynamic pieces. Now you can either attempt a database rebuild, or just go through the added area and select each piece to add them to their respective database. One other thing. I find that TSRE does not like new routes unless they already have a track and road database at first. Eric may have fixed this in more recent releases but I am using an older version that works best for me of TSRE. So I go to the start tile and pic some location a distance from the legitimate route and place one piece of track and one piece of road. Then move them just a teeny bit below the surface and leave them in place. This causes essentially empty database files to be created.

Jerry

[landnrailroader]

Colleague,

Another issue that comes up is how to digitize multiple tracks. Usually we are concerned with a passing track, double track, or a small yard. What I do is digitize the straightest route through a turnout, turnouts, or a rat maze. It is also possible to digitize the center between two tracks but I find it is better to digitize the straightest route and then lay the rest relative to it. For one thing we don't have an infinate number of possible track pieces and unless we use M-tracks, the spacing is nowhere near correct. Most yard tracks will be on 13 ft. centers and on the mainline, where passing speeds are higher, this may increase to 15, sometimes more feet.

But you just started a route that includes the NS mainline through the Carolinas. Nice single track but then @#$%^&* the heck is this, it looks like a wye turnout in the mainline. Well guess what, it is, you just came to a #20 equilateral turnout. These are relatively rare although there are also #24 turnouts or even longer. In the #20 equilateral, each track deviates from the center line of the incoming track on a factor of 1 in 40. A #20 turnout would normally be 1 in 20, i.e. measure 20 feet past the frog point & the inner rails should be 1 foot apart but in the equilateral they are 1 foot apart at 40 feet and this turnout is good for 80mph although speed is usually set at 50-60 max. just to reduce wear & tear.

We don't have such a thing, but I have been tinkering with M-track parts and such trying to create one. An advantage is that there is no left or right to a equilateral turnout, but then again, I only have so much hair to pull at 87. (Saturday, Jan. 24)

Jerry Sullivan

[landnrailroader]

Colleagues,

When digitizing tangent track off of Google Earth, frequently there is optical abberrations which make the tangent appear to be wiggly. If you also have a track chart you can determine if the stretch in question is really tangent or not. In digitizing the Carrollton RR, which is a branch off of the L&N Shortline from Worthville, KY to Carrollton (operated as a separate railroad but wholly owned now by CSX) such a situation is encountered where the road forks with the right branch running NE, parallel to the Ohio R. for a few miles to service many industries. The aerial image is a bit wiggly although the line is straight. So the thin red line is a measurement line drawn from the track center at this point, several hundred meters to the track center at the start of the next curve. Now I will extend the yellow digitized track to as precisely as possible follow the red line.

Jerry TANGENT.jpg