Message Areas - "DEM level 1 and 3 merging problems… "

Train-Sim.Com Message Areas

"DEM level 1 and 3 merging problems..."

Email this topic to a friend
Printer-friendly version of this topic
Bookmark this topic (Registered users only)
Previous Topic | Next Topic Conferences MSTS Route Design (Protected) Original message
wbgx (24 posts)
Dec-11-01, 05:32 PM (EST) "DEM level 1 and 3 merging problems..." Hello, I've lost quite a bit of hair attempting to get level 1 and level 3 DEM's to merge correctly. I've quickly searched this forum for an answer, but have not had sucess. Can level 2 and level 3 DEM's correctly merged without the big differences in elevations?Boomer Alert Edit | Reply | Reply With Quote | Top
Table of Contents

Subject Author Message Date ID RE: DEM level 1 and 3 merging probl... LawrenceKellie Dec-11-01 1 RE: DEM level 1 and 3 merging probl... wbgx Dec-11-01 2 RE: DEM level 1 and 3 merging probl... LawrenceKellie Dec-11-01 3 RE: DEM level 1 and 3 merging probl... wbgx Dec-11-01 4 RE: DEM level 1 and 3 merging probl... landnrailroader Dec-11-01 5 RE: DEM level 1 and 3 merging probl... LawrenceKellie Dec-11-01 6 RE: DEM level 1 and 3 merging probl... JohnCS Dec-11-01 7 RE: DEM level 1 and 3 merging probl... wbgx Dec-11-01 8 RE: DEM level 1 and 3 merging probl... wbgx Dec-11-01 9 RE: DEM level 1 and 3 merging probl... LawrenceKellie Dec-11-01 11 RE: DEM level 1 and 3 merging probl... LawrenceKellie Dec-11-01 10 RE: DEM level 1 and 3 merging probl... wbgx Dec-11-01 12

Lobby | Topics | Previous Topic | Next Topic
Messages in this topic
LawrenceKellie (923 posts)
Dec-11-01, 06:09 PM (EST) 1. "RE: DEM level 1 and 3 merging problems..." I'm sorry, but I am not clear on what DEM level 1 and 3 means. Could you provide a little more information? Are these DTED, GLOP...?Lawrence Remove | Alert Edit | Reply | Reply With Quote | Top
wbgx (24 posts)
Dec-11-01, 06:33 PM (EST) 2. "RE: DEM level 1 and 3 merging problems..." Lawrence, I appreciate your quick reply. I have downloaded DEM's from the gisdatadepot for my route. I don't remember the exact details, but they are the 24k free DEM's. I somehow successfully merged the subject DEM'S previously, but don't know how I did it now (very early in the morning after many failed attempts it finally works). I noticed in the gisdatadepot file area that several of the text files with the DEM download had a bunch of descriptions or spec's on the file, ie, resolution, and other information that didn't mean much to me. However, noticed that the spec's differed in the description of 'level 1' or 'level 3'. You may know what Iam describing but if I remember correctly the level 1 data when merged with the level 3 data had extremly high elevations at the DEM boundary's. I have followed the instructions in MircoDEM about the merging of 10 and 30 meter DEMs thinking that would apply to this case. Like I said eariler, I have successfully completed the XYZ file into my route folder, but have on idea how I completed the process. Now I need another DEM on the edge of my exsisting data and have the problem of getting it all to work with out the extreme high/low elevations at the DEM boundary planesI hope my description is enough for you to understand the problem Iam having with merging the different DEM's. Thanks in advance for any help you can give me.Boomer Remove | Alert Edit | Reply | Reply With Quote | Top
LawrenceKellie (923 posts)
Dec-11-01, 08:01 PM (EST) 3. "RE: DEM level 1 and 3 merging problems..." Ok. I see what you mean, and before today, I have never seen this before nor have I known that it makes a difference. In fact, I have no idea if it is even important for our usage.Anyone else?What about the person who posted the other day about writing a DEM tutorial (for free!).Lawrence Remove | Alert Edit | Reply | Reply With Quote | Top
wbgx (24 posts)
Dec-11-01, 08:19 PM (EST) 4. "RE: DEM level 1 and 3 merging problems..." It's crazy... What I have done is add DEM data to the adjacent data that looked fine. I first added the tiles in the route goemetry editor and then used the data in a XYZ format in the DEM-EX program. The results are that the area that needed the DEM data now has the tile generated (unfortunately, flat without terrian details) and lower than existing grade by, well down to sea level. Existing grade in my case is about 2100 meters. One thing thats goofy is that not all of the new tiles exhibit these characterstic, some of the new tiles are at the correct grade. Maybe I got a messed up DEM? I don't think so, but...Boomer Remove | Alert Edit | Reply | Reply With Quote | Top
landnrailroader (67 posts)
Dec-11-01, 08:29 PM (EST) 5. "RE: DEM level 1 and 3 merging problems..." I have noticed that these level values also are related to the resolution of the DEM.At the risk of having all kinds of slings & arrows, and rocks thrown my way, I am going to plunge ahead and describe how I got it to work. So far, XYZmerge does not work for me. Last night it ran over 10 hours and completed but did not ever produce an output file. The instructions from Lawrence' SD40 site indicated that I should go ahead and mix 10 & 30m DEMS and I think that was the problem. If one attempts to merge 10m DEMs, one soon runs out of available column & row space in MicroDem, as the help file indicates. About a month ago, I discovered a method that works for me, after which I use a suggestion of Lawrence's that also works, but not too well in one situation of mine.Download the desired DEMS from whereever you can get them, preferably for free. Each DEM has a range value, i.e. from one corner to the other, which is listed in the text file that goes with each DEM. If there are both 10m & 30m DEMS available, download only the 30m DEMS (smaller file size is the giveaway here), otherwise download whatever is available. It helps me to layout a grid, free-hand, and list the coordinate lat/long values on the edges, so I can list the DEMS in their proper relationship in the grid. That way, if I miss one, I know what to look for to complete my merge.Decompress the GZ files for the 30m DEMS, thus creating <name>.DEM files, where I choose the name of the Quad for the <name> of the file. Then resample the 10m DEMS, and when the screen appears giving the rows & columns, change these values to 1/3 of the value given rounded up. The process is explained in the help file and works very well if you follow those instructions. It might also be possible to use the "thin" command, but I have not tried this method. As the files are resampled, to reduce the density, give the resulting DEM files the naming convention as above.Now, using the instructions Lawrence has provided on several occasions, merge the DEM files into one DEM. I have found that if you try to add another DEM file on the edge of existing scenery, it will probably mess up existing scenery, so I found it was better to just use DEMEX on the whole route again (and again and again if necessary until it looks right). Convert the merged DEM into an XYZ file, and run DEMEX with it. Yes, you will mess up that nice work you already have done, but then it is a good idea to be sure you have all you need first, and in my case, I found that the second time through, using the "y" key went very fast, and I also knew where NOT TO USE IT, like at bridges and tunnels. I have put in track on 137 miles of mainline, and am completing the associated coal branches.Now, about SeamFill. This works great on relatively flat, or flat scenery, where "ditches" are smoothed out to the general contours of surrounding areas, but it doesn't cut it at all in mountainous terrain. I found that I had a ditch at a critical point on Blue Ridge, where the Clinchfield loops are (the loops are right on the juncture of 4 quads) and it was about 400 ft. deep. When I ran SeamFill, the result was a "spine" about 400 ft. above the surrounding terrain, so in this instance, I will manually correct the problem, which will be easy when I lay the track at the correct elevation.I did find that further down in SC, where a ditch occurred in relatively flat terrain, SeamFil handled it very well.Therefore, when the little map pops up and shows you the result of creating your massive merged DEM (or if you load it as a DEM file into MicroDem) note any streaks, usually running E-W that are in it. These will be on DEM boundaries. By referring to the GRID I mentioned making above, you can determine the precise lat/long of the centerlines of these ditches. Go into Geometry Extractor, and load the route and the Quad tree associated with it. Zoom into the route and put the cursor on the lat/long of one of the ditches. Note the tile names that appear in the header and jot these down. You will need the tiles along the C/L and one tile either side of the C/L. Create a directory tree on your hard-drive somewhere. I used D:\routes\blueridge\tiles and copy or move, the tile files (I copied all files with the same name, but evidently the ".t" file is the critical one) into the directory you created. Then run SeamFill on this directory, or group of tile files. The advantage to this, is that it runs faster, and it allows you to select the areas where SeamFill will probably do fine. In any questionable areas, you could keep a copy of the tile files in some other directory until you are satisfied with the result. If you cannot get a satisfactory result, just restore the offending files.J. H. Sullivan, P.E.
retired from CSX(Old railroaders never quit, they just get recycled - hence I'll be returning to work 1/2002 working through a Temp service.) Remove | Alert Edit | Reply | Reply With Quote | Top
LawrenceKellie (923 posts)
Dec-11-01, 08:35 PM (EST) 6. "RE: DEM level 1 and 3 merging problems..." LAST EDITED ON Dec-11-01 AT 08:36 PM (EST)Be very careful uncompressing GZ or TAR files with WinZip. There is a specific setting related to GZ/TAR files that MUST be set correctly or the data will be "butchered."The lastest version of MicroDem has a menu choice to uncompress files, and it consequently uses the correct programs--not WinZip. If you are going to uncompress the GZ/TAR files, as is suggested in the previous post, I would certainly encourage you to use MicroDem to do it.Lawrenceps: Just had a thought. Did you use WinZip instead of letting MicroDem uncompress it? Could that have been a factor? lrkpps: Just remembered this. The format of SDTS DEM was changed as of October 1, 2001. Could you possibly have a mix of the two types? Could that be a factor? lrk Remove | Alert Edit | Reply | Reply With Quote | Top
JohnCS (203 posts)
Dec-11-01, 08:54 PM (EST) 7. "RE: DEM level 1 and 3 merging problems..." When testing the Water Level route for someone I encountered problems with microdem merging dems and mis-scaling some of the terrain for one or two quads of the merged file (ie elevations which should have been 550 meters were 55 meters). Not sure why this happens.In that topic I posted a solution which worked for me. Don't recall what it was, but I think it involved resampling to different units. Search the forums on 'water level' and you should find it - its probably archived now. Remove | Alert Edit | Reply | Reply With Quote | Top
wbgx (24 posts)
Dec-11-01, 09:39 PM (EST) 8. "RE: DEM level 1 and 3 merging problems..." Thanks for everyones help. I reinstalled my route geographical data and got everything to somewhat work. First I merged the new DEM to the existing adjacent DEM (just one DEM not all 16 in this particular UTM zone). Then I DEM-EXed these (XYZformat)into the route. This caused some previous data to overwriten and leave voids in the geography. To fix this I re-DEM-EXed my original merged 16 DEMs (XYZ format) to the route. The results are better than before, but I still cannot achieve to 'pickup' the tiles north of my existing end point. I know I have the DEM data. I know the geometry editor has the tiles populated. So I wonder if my original geometry tiles have established boundaries which can not be exceeded? Maybe a goof idea, but I can not see what else would cause a failure of tiles to be generated.As you can see, from the original problem everything has evolved quite a bit. I plan to use the seamfil program next, but not until after I study the reason for the failure to add additional tiles.I originally had what I call level 1 and level 3 DEMs and after I put them together the results where huge jumps in the elevation of the terrain. I thought I merged decimeter and meter data together, but all where in meter units. In the route I would run into massive edges of depth or height where the change in the DEM's occured. Changes in magnitude of 10.Thanks your help and suggestions. I can imagine how I could get as far as I have without the sugestions and help you provide.Boomer Remove | Alert Edit | Reply | Reply With Quote | Top
wbgx (24 posts)
Dec-11-01, 10:10 PM (EST) 9. "RE: DEM level 1 and 3 merging problems..." Mr. Sullivan, I like the idea of using seamfil on only the tiles required as it now takes considerable time to run though the 1347 tiles in my route. I'll take advantage of your suggestion.Mr. Kellie (Lawrence), to the best of my knowledge I didn't use the WINzip program to use the DEM files. I simply loaded them within the MicroDem program. I don't know if I got a different format of data, ie, SDTS DEM.JohnCS, I searched the forum for the 'water level' tread or subject as you suggested and did not have any sucess. Most likely that information is archived.Thanks to you guys and others that contribute to this forum for giving MSTS the real potential to be a rewarding experince for everyone.Boomer Remove | Alert Edit | Reply | Reply With Quote | Top
LawrenceKellie (923 posts)
Dec-11-01, 10:46 PM (EST) 11. "RE: DEM level 1 and 3 merging problems..." ::sigh::Why the Mr. Kellie?If you look, my comment was in response to landnrailroader's post not yours. He and I have exchanged several e-mail over the last several days trying to get his problem fixed.It occurred to me when I read his post and he mentioned uncompressing the files, that that may be *his* problem, not yours.So, between that and the Mr. Kellie, I have no idea what is going on.Lawrence Remove | Alert Edit | Reply | Reply With Quote | Top
LawrenceKellie (923 posts)
Dec-11-01, 10:43 PM (EST) 10. "RE: DEM level 1 and 3 merging problems..." Definition of Level 1, 2, and 3:USGS DEM - Technical Details: Classification LevelsRoot-mean-square error (RMSE) is a measure of how closely a data set matches the actual world Level-1 DEMs reserved for 7.5-minute DEMs created by scanning NAPP photography vertical RMSE of 7 meters is the desired accuracy, 15 meters is the maximum permitted. Level-2 DEMs have been processed or smoothed for consistency and edited to remove identifiable systematic errors, derived from hypsographic and hydrographic data digitizing RMSE of one-half contour interval is the maximum permitted. Level-3 DEMs are derived from DLG data from both hypsography (contours, spot elevations) and hydrography (lakes, shorelines, drainage) RMSE of one-third of the contour interval is the maximum permitted. Sometimes you will encounter both level 1 and 2 DEMs for the same area; the higher level number is much better data and should be used Lawrence Remove | Alert Edit | Reply | Reply With Quote | Top
wbgx (24 posts)
Dec-11-01, 10:58 PM (EST) 12. "RE: DEM level 1 and 3 merging problems..." Well I get to thank you now! ;)Boomer Remove | Alert Edit | Reply | Reply With Quote | Top

Lock | Archive | Remove

Lobby | Topics | Previous Topic | Next Topic