New Canadian elevation data!

[Holger 0]

Hi all:just received news about a new Canadian government site that contains, among other things, previously unavailable elevation data, at two resolutions: 1:250,000 (3 arcsec) and 1:50,000 (0.75 arcsec). The lower-resolution data covers all of Canada and the higher resolution is currently available for selected areas of most provinces and all of B.C. (the latter is hard to believe as our provincial government charges several hundred dollars per 1:50,000 digital mapsheet!).This is the website address: http://www.geobase.caHere's the current coverage of the two resolutions (top 3 arcsec, bottom 0.75 arcsec):http://forums.avsim.net/user_files/49234.jpghttp://forums.avsim.net/user_files/49235.jpgHow do the data compare to SRTM 3-arcsec data? Well, the 1:250,000 data are of similar resolution but don't contain any missing areas and extend north of N60. As for the high-res data, the difference is quite amazing, at least when viewed with 3DEM (http://www.visualizationsoftware.com/3dem.html). I haven't had a chance yet to compile a FS mesh as Microdem seems to struggle with the data format; shouldn't be a big deal though.Here are a few examples of the 0.75 arcsec data of British Columbia; compare the difference in detail and the lack of missing data in the CDED1 files:http://forums.avsim.net/user_files/49236.jpghttp://forums.avsim.net/user_files/49237.jpghttp://forums.avsim.net/user_files/49238.jpgPretty exciting, huh? Cheers,Holger

[Guest jkvato]

Looks like you've got work to do!Looking forward to what you can do with it ...

[Holger 0]

Me again,just completed a sightseeing trip of Glacier National Park in all its CDED1 LOD10 (38-m) splendor - no more weird peaks and slopes! Simply breathtaking!For now I'm using a workaround via 3DEM to get the data into MicroDEM for merging and processing. If anyone finds a way to import the files directly, please let us know.Happy meshing ;-)Cheers, Holger

[gaugi1 0]

Hi Holger,The joined picture compares elevations of the same territory with new Canadian Digital Elevation Data (3 sec/arc) and SRTM-3.The SRTM-3 source gives best results. But there is a good source to correct, with MIcrodem, the missing elevation in SRTM-3.SalutGilleshttp://forums.avsim.net/user_files/49298.jpg

[Holger 0]

Hi Gilles:thanks for that comparison and info!It seems like adding another image to the ones in my original post above has once again deleted the first set of images (I've asked the AVSIM staff about this before but they never gave me an answer). Anyway, here are some of the original images: top is 1:50,000 coverage, then 1:250,000 coverage, then a comparison of CDED1 1:50,000 data with SRTM data.http://forums.avsim.net/user_files/49302.jpghttp://forums.avsim.net/user_files/49308.jpghttp://forums.avsim.net/user_files/49306.jpgCheers, Holger

[Guest PAtom]

QUOTE http://library.usask.ca/data/geography/nrcan/cded.html :Data FormatCDED are recorded in USGS ASCII format.The CDED were produced using Australian National University Digital Elevation Models (DEM) software. These DEMs have the same characteristics as United States Geological Survey digital elevation models. They are written as ANSI Standard ASCII characters, recorded in IBM Standards fixed-block format.so perhaps it is possible to convert these data into bgl by using resample.exe?tom

[gaugi1 0]

Hi Holger,To complte my post, see the merging result in using CDED-3 and SRTM-3.SalutGilleshttp://forums.avsim.net/user_files/49316.jpg

[Holger 0]

Hi Tom:It would be nice if it were that easy ;-)Resample.exe requires binary 16-bit BSQ format (transformed from 16-bit integer elevation data, LSB first), as is stated in the MS Terrain SDK documents. The newer MicroDEM versions (7.01 etc.) can save a DEM as 16-bit BSQ but it has problems reading the CDED1 .dem files, at least for me.In MicroDEM 7.01 (build 14.59.19.1, 11/11/2003), I receive a range check error when trying to open a CDED1 1:50,000 .dem file.In MicroDEM 6.0 (build 14.29.13.3, 8/13/2002), I can load the file and the map window corners show the correct spatial extent but the header information is wrong (e.g, it states 901 rows instead of 1201). When trying to merge several tiles the result is a pretty funny jigsaw puzzle. Strangely, the same process seems to work fine with the 1:250,000 data.My work-around is this: load each .dem tile into 3DEM and save as geotiff. These can be read, merged, and subset correctly with MicroDEM 7.01 and saved as 16-bit BSQ ready for FS resample.exe. It would be nice to eliminate this extra step so if anyone has any ideas...Cheers, Holger

[Holger 0]

Hi Gilles:hey, image thief :~P ;-)Nice pictures - it's a bit hard to see (squint, squint) but it seems like there are less problems with the visible seams in merged areas; that's good news! So, when are you going to deliver northern Quebec and Baffin Island? :-boom Cheers, Holger

[gaugi1 0]

Hi Holger,I have made the northern Quebec, but in the Baie James region (in north-west) all the lakes are sunken and in northeastern region, all the lakes are raised. I have not uploaded it because it's not nice to fly. And too much work to correct.For the Baffin Island, i don't think of doing this territory.SalutGilles

[Holger 0]

Howdy,more screenshots of my first test CDED1 mesh available here: http://portal.fsgenesis.net/modules.php?se...=view_album.phpOne thing I noticed in another area (Garibaldi Peak) is that a higher TMVL level (20 or 21, instead of 19), while necessary for revealing all detail, also shows a regular pattern of "bumps". In my experience, this is typical for algorithms that interpolate surfaces from a regular grid of height data, particular if the final grid has more than one grid point between two known elevations. Usually, the processor would attempt to remove the "bumps" with a smoothing filter but it doesn't look like the people at NRCan have done this. Oh well, it's not obvious in most places and even with TMVL set at 19 the data still look much better than the SRTM data. Cheers, Holger

[Flyer Ken 0]

Holger, your images really inspired me to look into mesh design. I'd like to do some work on Vancover Island, where I live. I got some DEM files from the geobase site. I installed 3DEM and MicroDem 7. I know I have to merge DEM files to make a larger area, but I thought I'd start of with just one DEM file. I managed to open one of the DEM files in 3DEM and save it as a geotiff file, which opened sucesfully in MicroDem. The question is, what do I do now? You mentioned saving it in 16 bit BSQ format, but I can't figure out how to do that - is there a command in MircoDem?

[Holger 0]

Hi Ken,that's great - there can never be enough mesh designers ;-)As for Vancouver Island, I just started on the SE part that is covered by Harvey Janszen's new coastline/landclass replacement files; screenshots are here: http://www.flightsimnetwork.com/dcforum/DC...D21/9539.html#3It's a bit of a tricky area and requires working with the default LWM flatten files. My files probably won't extend west of W123.8 longitude so there's lots of VI left to cover but you can, of course, build mesh anywhere you like.I assume you are familiar with the basics of how mesh is displayed, particularly, that only parts fully inside a LOD grid cell (e.g., LOD10) will display in FS. In other words, if you make independent LOD10 mesh files out of each SDED1 .dem file you'll end up with gaps between the tiles. There's no step-by-step tutorial for mesh design out there but the site that comes closest to that would be Steve's: http://www.fs-traveler.com/welcome.shtml. In addition, you do want to read the FS2002 Scenery SDK (http://www.microsoft.com/games/flightsimulator/fs2002_downloads_sdk.asp ) for the basics. Finally, the MS TMFViewer utility is the most efficient tool for checking your mesh before installing in FS (part of the Custom Terrain Textures SDK on the same site). TMFViewer will show you the mesh extent displayed in FS, so you can easily detect gaps and other unintended problems.The function for saving in BSQ-16 format is via File > Save DEM > 16-bit BSQ (use -9999 for missing data if you're using my .inf file below as a template). If you don't see it in your version you'll have to download the latest build from the MicroDEM website (the versions are updated every few weeks).Below is a sample .inf file I used for my VI test mesh. Of course, you'll have to adjust it to fit your parameter values. Also, I suggest giving the .inf file the same name as the .dem file your compiling.[Destination]; LOD = {Auto, 0..n}; LOD = Auto tells the resampler to pick the closest LOD to the;source;cell size. Specifying a number will force an LOD. This can;be used for undersampling or oversampling.LOD = 10DestDir = "F:tempBC"DestBaseFileName = "mainld_s"UseSourceDimensions = 1[source]Type = ElevS16LSBSourceDir = "F:tempBC"SourceFile = "mainld_s.dem"NullCellValue = -9999Lat = 50.000000Lon = -124.0000000NumOfCellsPerLine = 10801NumOfLines = 4801CellXdimensionDeg = 0.000208333333333333CellYdimensionDeg = 0.000208333333333333ScaleinMeters = 1.0For the actual use of the .inf file with the MS Terrain SDK utilities, I've set up a small .bat file that looks like this:resample mainld_s.inftmfcompress mainld_s.tmf mainld_s_c.tmftmf2bgl mainld_s_c.tmf mainld_s_c.bglReplace mainld_s with the name of your .inf file, save as .bat and double-click to start (making sure that both the .inf and the final .dem file are in the indicated directory).Cheers, Holger

[Flyer Ken 0]

Hi Holger, thanks for your detailed answer. I'm going to try to get into it this week. I do indeed have Harvey's excellent land class and coast line scenery. That's the area that I'm most interested it. I'll slog on up the learning curve, and if I get to the point where I can do some useful work, I'll get back to you to make sure I don't duplicate anything you are working on.Thanks again for the help.