Jump to content

 
Photo

Elevation Data Collection

- - - - -

  • Please log in to reply
19 replies to this topic

#1
franciscocartographer

franciscocartographer

    Key Contributor

  • Validated Member
  • PipPipPip
  • 58 posts
  • Gender:Male
  • Location:Phoenix, Arizona
  • Interests:Maps, FC Barcelona, Cadiz CF, Phoenix Suns, Hiking
  • United States

I was wondering if is possibe to create higher resolution DEMs (>30 meters) from DLGs or DRGs.

I am doing elevation maps for a small area (600 acres) and 30m resolution DEMs are a bit too coarse.

I will be very grateful for any advice on this.
--------------------------------------
Francisco Jimenez, GISP
Senior GIS Analyst & Amateur Cartographer

My webpage

#2
Hans van der Maarel

Hans van der Maarel

    CartoTalk Editor-in-Chief

  • Admin
  • PipPipPipPipPipPipPip
  • 3,869 posts
  • Gender:Male
  • Location:The Netherlands
  • Interests:Cartography, GIS, history, popular science, music.
  • Netherlands

You could digitize the contour lines from a raster map, attach the correct elevation attribute to them and from there, generate a terrain model at whatever resolution you'd like. There's several software packages that can do that, I would personally go for FME or VNS, but I'm sure the ESRI products will be able to do this, Manifold probabely can and also Leveller.

Alternatively, have you considered using the NED? I can't recall off the top of my head what it's resolution is, but definately better than 30m.
Hans van der Maarel - Cartotalk Editor
Red Geographics
Email: hans@redgeographics.com / Twitter: @redgeographics

#3
Martin Gamache

Martin Gamache

    Ultimate Contributor

  • Associate Admin
  • PipPipPipPipPipPip
  • 980 posts
  • Gender:Male
  • Location:Washington DC
  • Interests:History of Topographic Cartography
    Topographic Mapping
    History of Relief Depiction
    Thematic Cartography
    Demographic Cartography
    Cartographic techniques, methods, and tools
    Orienteering
    Panoramic & Kite Photography
  • United States

Francisco,

If you have access to the contour and elevation points in vector formats I would use the Topogrid module (Arcinfo, Spatial or 3d analyst) and generate a 10m DEM. I might also use a lake and hydrography layer as additional inputs depending on the planned use of the DEM. This module is fairly robust and originates from ANUDEM. Its output have been well described and documented in numerous publications. I recommend the versions in Arc 9, they are very easy to use from the GUI. For purely cartographic work its output are very nice but leave out the hydro layers from your calculations.


mg

#4
franciscocartographer

franciscocartographer

    Key Contributor

  • Validated Member
  • PipPipPip
  • 58 posts
  • Gender:Male
  • Location:Phoenix, Arizona
  • Interests:Maps, FC Barcelona, Cadiz CF, Phoenix Suns, Hiking
  • United States

Thank you for both inputs.

My first thought was to do what Hans said about digitizing contours from DRG or other raster file and generating the elevation model using Spatial Analyst or IDRISI.

I have never tried the Topogrid module for ArcGIS but I will try it to model a 10 M DEM.

Thanks again!!
--------------------------------------
Francisco Jimenez, GISP
Senior GIS Analyst & Amateur Cartographer

My webpage

#5
Martin Gamache

Martin Gamache

    Ultimate Contributor

  • Associate Admin
  • PipPipPipPipPipPip
  • 980 posts
  • Gender:Male
  • Location:Washington DC
  • Interests:History of Topographic Cartography
    Topographic Mapping
    History of Relief Depiction
    Thematic Cartography
    Demographic Cartography
    Cartographic techniques, methods, and tools
    Orienteering
    Panoramic & Kite Photography
  • United States

Francisco,

Just to clarify I am essentially saying the same thing as Hans, just recommending a specific tool (topogrid) that I know to be very good for this purpose.

Topogrid is part of Spatial Analyst, you will find it in the toolbox under raaster to vector I think or soemthing similar. If you do not have a set of contour / spot elevation points digitized from the DRG you will need to digitize those. DO not use contours that were interpolated from the 30m DEM. But it sounded like you already had the DLG vector contours?

As for NED it is in plate carre and resolution is reported in arcseconds which depending on your latitude affects the projected resolution. But for outside the US and closer to the Equator it is SRTM data at approximately 90m and can be prone to errors on steep slopes. For areas in the US you can sometimes find 10m DEMs.

mg

#6
franciscocartographer

franciscocartographer

    Key Contributor

  • Validated Member
  • PipPipPip
  • 58 posts
  • Gender:Male
  • Location:Phoenix, Arizona
  • Interests:Maps, FC Barcelona, Cadiz CF, Phoenix Suns, Hiking
  • United States

Actually I do not have DLGs but I could download them in SDTS format. DRGs are easier to download but I will have to convert them to vector (digitize the contour lines).

I have the ARc GIS8.3 version with spatial analyst, so topogrid is not included. But I think IDRISI Kilimanjaro has a similar way (althought it does not use the ANUDEM Algoritm) of creating DEMs from contour lines.

Also 10m NEDs I think are only available for the North West of USA.
--------------------------------------
Francisco Jimenez, GISP
Senior GIS Analyst & Amateur Cartographer

My webpage

#7
vasile33

vasile33

    Newbie

  • Validated Member
  • Pip
  • 6 posts
  • Location:Bucharest
  • Romania
Francisco,

If you are using ArcGIS(ArcView)+Spatial Analyst you can interpolate the extracted vector contour lines and elevation spots using “Interpolate to raster” function. You may choose from 3 different algorithms. To semi-automatically extract the contour lines you can use ArcScan under the same environment (ArcGIS). A free alternative (both digitize and interpolate) is Blackart (www.terrainmap.com). Other free tools to interpolate contour lines are: Landserf (www.landserf.org), SAGA or DiGeM (http://www.saga-gis....tingen.de/html/ , http://www.geogr.uni...gem/index.html), Microdem (http://www.usna.edu/...te/microdem.htm). Another (commercial) option is Surfer from Golden Software (www.golden.com). The demo version will allow you to interpolate the contour lines into a grid using 12 different algorithms.

Another trick you can do is to resample and smooth the original 30 meters NED file. This is some how similar with the re-interpolation of the contour line extracted from NED but is faster and maybe looks more natural. This approach is not recommended if you intend to perform any spatial analysis over the DEM but may produce a good result if you want only a background for your map. At http://www.romaniadi...la.ro/test.html i put an comparison between the original NED file and the one obtained after applying “Spline Smooth” filter in Surfer. This function works without restrictions in the demo version of Surfer. The data in the screenshot covers a area of 4.5 square miles.

Vasile

#8
Martin Gamache

Martin Gamache

    Ultimate Contributor

  • Associate Admin
  • PipPipPipPipPipPip
  • 980 posts
  • Gender:Male
  • Location:Washington DC
  • Interests:History of Topographic Cartography
    Topographic Mapping
    History of Relief Depiction
    Thematic Cartography
    Demographic Cartography
    Cartographic techniques, methods, and tools
    Orienteering
    Panoramic & Kite Photography
  • United States

IMO Blackart is still too buggy for production work. Good theoretical foundation but very inconsistent output, that is if you manage to get an output.

mg

#9
vasile33

vasile33

    Newbie

  • Validated Member
  • Pip
  • 6 posts
  • Location:Bucharest
  • Romania
Yes, Blackart is not the friendliest or the most stable solution. I’m using Blackart only for interpolation. The results are quite good but the processing time is huge. Usually I let Blackart to run during the night :)

Other good piece of software is SurGe (http://www.geocities...r/surgemain.htm). The interface is quite messy but the software is very flexible and robust. The shareware version have some limitations but the full version cost only 20$.

Vasile

#10
Martin Gamache

Martin Gamache

    Ultimate Contributor

  • Associate Admin
  • PipPipPipPipPipPip
  • 980 posts
  • Gender:Male
  • Location:Washington DC
  • Interests:History of Topographic Cartography
    Topographic Mapping
    History of Relief Depiction
    Thematic Cartography
    Demographic Cartography
    Cartographic techniques, methods, and tools
    Orienteering
    Panoramic & Kite Photography
  • United States

Vasile

Of all the free utilities you mentioned I think SAGA is the most promising and has been the most useful to me in my work. I used both Surge and Landserf at some point and they seemed more like research tools and still a bit buggy. Microdem can be a good diagnosis tool with problematic data. I have not used SAGA in over a year but last I checked they were developing alot of modules for it and for free software it has gotten quite powerful. Reminds me a bit of IDRISI in some ways.

I have not used Surfer beyond playing around with a 2 year old version. Although my impression was that it was fairly limited in the input formats especially if working with contour data and was more designed to work with point data, which it seems to do quite well.

mg

#11
vasile33

vasile33

    Newbie

  • Validated Member
  • Pip
  • 6 posts
  • Location:Bucharest
  • Romania
Martin,

All the software I mention (except Surfer) represents results of scientific work. SAGA is an interesting project lead by Prof. Olaf Conrad from the University of Goettingen and is constructed on the roots of DiGeM. SAGA is very powerful when you do raster analysis and has a lot of functions for geomorphological analysis using well documented algorithms. More difficult is to work with vectors and to produce hi-quality cartographic layouts. The situation is almost similar with Landserf, Microdem or TAS (http://www.sed.manch...y/research/tas/).

In my first post I didn’t mention GRASS. Everybody knows about GRASS but just (relative) small community is using. In GRASS there is implemented a very interesting gridding method (s.surf.rst) called RST (Regularized Spline with Tension) developed by Helena Mitasova and Lubos Mitas. RST is a multivariate interpolation method some how similar with the one you’ll find in expensive geo-statistical package (like Geostatistical Analyst for ArcGIS) and the output grid don’t present artifacts like terracing, pits or, more important, the banding effect of the elevation values around the contours, specific to other smooth interpolation methods (like Kriging).

Vasile

#12
franciscocartographer

franciscocartographer

    Key Contributor

  • Validated Member
  • PipPipPip
  • 58 posts
  • Gender:Male
  • Location:Phoenix, Arizona
  • Interests:Maps, FC Barcelona, Cadiz CF, Phoenix Suns, Hiking
  • United States

Vasile,

I have used SAGA and it has very powerful raster capabilities. On the other hand it still has some 'bug' problems and the user group community is not as big.

I have never used GRASS but I am thinking on downloading it. I understand that it is used by many academic, government, and commercial entities so the user community is quite large.

Could you compare GRASS with other GIS Applications? I have used ArcGIS, SAGA, IDRISI, SURFER (not a GIS), and MapInfo.

Thank you in advance.
--------------------------------------
Francisco Jimenez, GISP
Senior GIS Analyst & Amateur Cartographer

My webpage

#13
vasile33

vasile33

    Newbie

  • Validated Member
  • Pip
  • 6 posts
  • Location:Bucharest
  • Romania
Francisco,


GRASS is a full GIS product. SAGA and Surfer implement only a limited set of GIS functions. The GRASS community is relatively large but some how exclusivist. That’s because GRASS is not so easy to install and run and the learning curve is a little bit steep. There is some documentation on GRASS website (http://grass.itc.it/gdp/) but is some how focused on programming aspects and not on describing the application philosophy. Maybe the “Open Source GIS A GRASS GIS Approach” book provide a better image of the GRASS software. The interface is another disadvantage for GRASS. If you are not migrating from ArcInfo command line environment it will be quite a challenge to work with GRASS tcltck menus and command lines. That’s why the normal users may give up on GRASS. With a little bit work you can configure GRASS to work with a more decent graphical interface via QGIS or JavaGRASS (http://grass.itc.it/...enshots/gui.php). GRASS is very powerful when you do raster analysis or image processing; it uses open and scientifically proved algorithms and is very fast. The latest versions include a good vector engine.
GRASS runs on a variety of platforms but also require some external libraries to make it run. On Windows you’ll need Cygwin.

As a conclusion, I’ll say that GRASS can perform all the data manipulation and geoprocessing tasks (maybe faster and cleaner) that ArcGIS or MapInfo do. Much more difficult is to produce a high-quality map inside GRASS. I usually use GRASS only for analysis or batch data manipulation and I make my maps with ArcGIS and DTP software.


Regards,
Vasile

#14
franciscocartographer

franciscocartographer

    Key Contributor

  • Validated Member
  • PipPipPip
  • 58 posts
  • Gender:Male
  • Location:Phoenix, Arizona
  • Interests:Maps, FC Barcelona, Cadiz CF, Phoenix Suns, Hiking
  • United States

Thanks for the input, Vasile.
--------------------------------------
Francisco Jimenez, GISP
Senior GIS Analyst & Amateur Cartographer

My webpage

#15
dylan

dylan

    Contributor

  • Validated Member
  • PipPip
  • 14 posts
  • Interests:soil science, open source software, map making and analysis
  • United States

Stumbled onto this thread, and thought that I would post my take on what has been discussed so far.

1. Smoothing and or re-interpolation of elevation data. As has been said to some degree before, it really depends what you want to do with the data. If it is only a backdrop, smoothing via a neigborhood function (average) would be more than adequate. Attempting to re-interpolate a new DEM from a topo map (DLG/DRG) would be an excercise in futility- this is how the lower quality NED data was produced in the first place, and any such product would be likely be inferiour to even the NED data.

2. Interpolation in general. After an extensive literature review for similar research, interpolation from contours or points is probably best done via aundem or v.surf.rst (GRASS). Both methods have their own good and weak points, I would recommend reading up on either of them. In general tradiational geostatistical methods will not produce a natural looking or hydroligcally correct DEM. But since the quality of the input data is the most important factor, your mileage may vary. Anudem can potentially produce a better hydrologically correct DEM, but v.surf.rst is more flexible.

3. GRASS. The user community is large and international with a very active mailing list. Development is onoing, and quite frequent- in the CVS version you can expect changes every couple of days. In addition, the lead developers are very open to new users. However, users are expected to be willing to learn about GRASS, and more importantly understand the subtlties of what a GIS is all about. This can be a substantial stumbling block at first, especially to those who are used to pointing and clicking, and not always understanding. The documentation is still a little terse, but there is active work on improving it- somehting that I am greatly looking forward to. The production of high quality output from GRASS can be accomplished with GMT. I am currently involved in the development of some tools to facillitate this process.

4. the problem at hand. Another technique for getting a better 'looking' DEM at such a large geographic scale would be to use an interpolatoin technique on the raster itself to re-sample the 30meter DEM to something like a 5 to 10 meter DEM. in GRASS this could be accomplished with r.resamp.rst - using the afore metioned Regularized Splines with Tension algorithm.

my 2 cents,




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

-->