Creating terrain heightmaps

  • trau dich einfach auf ok. Und dann gibt es noch ein Eigenschaftsmenü. ;)

    sollte aber auch das gleiche wie in qgis stehen.


    Sorry, I cannot check your stuff in the bath tube. And I am on the way to the aerobatic nationals. So no time. There will be an article in the german FS Magazin in one of the next issues where I describe how I do elevation mesh. So stay tuned. ;)

    Wish for Aerofly FS 2:

    - Flightpath recording on hard drive and replay in sim from different view points

    - Smoke for aerobatic planes

    - Multiplayer or at least watching other people flying sitting on ground or inside tower

  • Sorry, I cannot check your stuff in the bath tube.

    Why not? :)

    And I am on the way to the aerobatic nationals.

    All the best! :thumbup:


    I tried it again with another source: http://viewfinderpanoramas.org/dem3.html#alps%E2%80%8B


    FS 2 standard mesh for Mauritius:


    My creation:


    Still not very impressive, but better than the standard.


    It seems, that there is something wrong with the other source: http://srtm.csi.cgiar.org/srtmdata/


    Cheers, Luca.

  • I see you got it working, its a 30 metre mesh so improvements will be modest.


    Just for some info for you and others needing help.

    I think your source is fine just your files are wrong, just use the tfw file that came with the download, its perfect.

    here are mine


    Contents of srtm_67_19.FTW


    0.0008333333 < pixel resolution, yours should be same as we both using 6000x6000 file from the same source, check your hdr & files its in there.

    0.0000000000

    0.0000000000

    -0.0008333333 <pixel resolution, yours should be same as we both using 6000x6000 file from the same source, check your hdr files its in there.

    150.0000000000 < Change Coords to suit your location again can be found in hdr file an TFW file

    -29.9999995157 < Change Coords to suit your location again can be found in hdr file an TFW file



    My mesh_conv.tmc


    <[file][][]

    <[tmcolormap_regions][][]

    <[string] [folder_source_files][./input_aerial_images/]>

    <[bool] [write_images_with_mask][false]> // global flag -> if this is true, no image is created

    // if a mask is required. if this is false you can still

    // disable mask images for a selected level, see below

    // images with mask are always 'problematic'. rule of thum

    // is to create an area as large as possible


    <[bool] [write_ttc_files][false]>

    <[bool] [do_heightmaps][true]>

    <[string8][folder_destination_ttc][./scenery/images/]>

    <[string8][folder_destination_heightmaps][./scenery/images/]>

    <[bool] [always_overwrite][true]>



    <[list][region_list][]


    <[tmheightmap_region][element][0]

    <[uint32] [level] [11]>

    <[vector2_float64] [lonlat_min] [149.9995833 -29.9995828]>

    <[vector2_float64] [lonlat_max] [155.0004167 -35.0004162]>

    <[bool] [write_images_with_mask][false]> // do not create images that would require a mask

    >

    >

    >

    >


    Not only did I read QWERTY42's tutorial but also another by crispy136

    https://flight-sim.org/filebas…mesh-guide-for-australia/


    You will notice I only used Level 11 like crispy136 did.

    This makes sense in my mind as you can still see the Level 11 mesh at higher levels any way.

    Similar to how you can see the default Level 7 ish mesh at levels 8 to 15 or so.


    So making meshs for levels higher than level 11 is kind of redundant.

    Please anyone correct me if Im wrong in this assumption

  • So making meshs for levels higher than level 11 is kind of redundant.

    Please anyone correct me if Im wrong in this assumption

    It's only useful to convert beyond level 11 if your starting mesh data is higher in resolution than an AeroFly level 11 tile.


    In AeroFly & geoconvert, the 'levels' are really resolution levels. As you get closer to the ground, you need a higher resolution (level) to keep things looking good. As you get closer to the terrain, you'll see level 9 data, then level 10 data, then level 11 data and so on. If you're right on the ground or very close to it, you'll see level 15 data, assuming you actually have level 15 tiles converted.


    Each 'level' in AeroFly has a certain resolution associated with it (this resolution actually varies a bit with latitude because of how the tiles are mapped to the spherical Earth). Your downloaded geotiff source data also has a certain resolution. If your source data isn't at least as high in resolution as the AeroFly level you're converting to, then it is indeed a waste of time because it won't gain you anything by converting it. It's like zooming in further on a photo when you're already seeing the individual pixels in it. It will still convert just fine, it's just that each successive level beyond the resolution maximum of your source geotiff won't have any more detail than the one before it. Hope that makes sense.


    Also, here are the detailed values of that resolution limit, for each level and varying with latitude, in meters per pixel: resolution vs level vs latitude link

  • "Also, here are the detailed values of that resolution limit, for each level and varying with latitude, in meters per pixel: resolution vs level vs latitude link"


    Thank you @querty42 . I was always a bit curious about people giving the meter per pixel value without the latitude in mind. The consequence is also, in a 2024x2024 tile the lat resolution is different to the lon resolution. :)


    Also the resources of geotiffs can have different sizes and thus different resolutions. I had at 49 degree a perfect square, at 50 degree the height was twice of the width. You have to calculate your output maximum level depending on all this informations, if you make a good job. :)

    Wish for Aerofly FS 2:

    - Flightpath recording on hard drive and replay in sim from different view points

    - Smoke for aerobatic planes

    - Multiplayer or at least watching other people flying sitting on ground or inside tower

    Edited 3 times, last by flightxtreme ().

  • "Also, here are the detailed values of that resolution limit, for each level and varying with latitude, in meters per pixel: resolution vs level vs latitude link"


    Thank you @querty42 . I was always a bit curious about people giving the meter per pixel value without the latitude in mind. The consequence is also, in a 2024x2024 tile the lat resolution is different to the lon resolution. :)

    Yep, that's correct! In those plots and the table at that link, the numbers I've shown are the most detailed resolutions for the entire tile, meaning the lowest value in units of meters/pixel for each. This takes into account the different lat and lon resolutions in a given tile. I presented it that way because it lets you know what imagery resolution you need in order to fully take advantage of a certain AeroFly level.

  • Just for some info for you and others needing help.

    I think your source is fine just your files are wrong, just use the tfw file that came with the download, its perfect.

    Thank you very much for your help. :thumbup:Unfortunately, I can't find anything wrong in my files. Maybe, we'll keep it as it is and I use the other source, which gives me fine results.


    Not only did I read QWERTY42's tutorial but also another by crispy136

    https://flight-sim.org/filebas…mesh-guide-for-australia/

    Good work!!! :)


    Also, here are the detailed values of that resolution limit, for each level and varying with latitude, in meters per pixel: resolution vs level vs latitude link

    Great graphics and tables. Is there a rule of thumb for laymen which levels are correct for 5 m, 10 m and 30 m?

  • I was talking about your try with the tiles from J. de Ferranti (viewfinderpanoramas.org).

    Sorry, I got you wrong. I will do the alps mesh and made a positive test with N47e015. Also, SF40 for Mauritius was fine.


    What are the right levels for the meshs?

    • querty42 tells in his tutorial, that we should convert the levels 12-14, when we work on a 10 m mesh.
    • abbiati50 created his italy mesh (19 m) with the levels 8-11 (incl. 10)
    • Taburets canary islands (5 m) are done with 8, 9, 11, 12 (excl. 10)
    • OrbX uses for its meshes the levels 10 and 11.
    • Vogel 69's Martinique (1 m) is done with 11, 13, 14.
    • Rodeo's Washington (? m): Level 10.
    • Aerofly produced the levels 7, 10, 11 for the meshs with different meters/pixel.

    I can't see any system in the above examples. I would like to create a 10 m mesh for the Alps. Can anyone help me? Which levels should I produce? :/

  • Sorry, I got you wrong. I will do the alps mesh and made a positive test with N47e015. Also, SF40 for Mauritius was fine.

    Ok, N47 is the north of the Alps without see. As said in the other thread you can have water fall for the southern Alps if you don't clean the DEM from J. de Ferranti.


    That's also the case with SRTM tiles from USGS but only with few void pixel in comparison with the ones from de Ferranti where the entire oversea area has void value.


    Both can be easily fixed but if the area you want to build is available on viewpathfinder.org (it's the case for the whole Alps) it's better to use files from J. de Ferranti because from what I've seen the quality of the DEM is much better (even both are based on SRTM 1 arc).


    What are the right levels for the meshs?

    • querty42 tells in his tutorial, that we should convert the levels 12-14, when we work on a 10 m mesh.
    • abbiati50 created his italy mesh (19 m) with the levels 8-11 (incl. 10)
    • Taburets canary islands (5 m) are done with 8, 9, 11, 12 (excl. 10)
    • OrbX uses for its meshes the levels 10 and 11.
    • Vogel 69's Martinique (1 m) is done with 11, 13, 14.
    • Rodeo's Washington (? m): Level 10.
    • Aerofly produced the levels 7, 10, 11 for the meshs with different meters/pixel.

    I can't see any system in the above examples. I would like to create a 10 m mesh for the Alps. Can anyone help me? Which levels should I produce? :/

    Nice summary !


    For your 10m mesh I would say level 9-11 or maybe 8-11 knowing that the level 11 in AFS has a resolution of 9.54m / pixel.


    Can I ask you what is this DEM of 10m you have for the Alps ?

  • As said in the other thread you can have water fall for the southern Alps if you don't clean the DEM from J. de Ferranti.

    Thank you for the advice. It's great, that you found such an easy solution. Do you think I should clean all tiles with QGIS' as a precaution?



    viewpathfinder.org

    I don't find this source. Maybe, you mean http://viewfinderpanoramas.org/dem3.html#hgt? Here are files from J. de Ferranti.



    For your 10m mesh I would say level 9-11 or maybe 8-11 knowing that the level 11 in AFS has a resolution of 9.54m / pixel.

    Thank you very much! Is it the same system as the ortho files?



    Can I ask you what is this DEM of 10m you have for the Alps ?

    It's from here: http://viewfinderpanoramas.org/dem3.html#alps


    Isn't 1 " equivalent to 10 m? ?(

  • Oops I was talking about viewfinderpanoramas.org, I don't know why I wrote viewpathfinder.org ! :)


    SRTM 1" means resolution of 1 arc second, so it means ~30 meters (at the equator), not 10 meters.


    I really don't think there is a free DEM with a resolution of 10 meters which covers all the Alps.


    But the 30m DEM of J. de Ferranti are very good even if the resolution is not very high. And from I've seen there are already cleaned of void pixels, except for the oversea area (so only 3 tiles should be concerned: N43005, N43006, N43007).

  • As counter-intuitive as it may read, a very high resolution mesh for mountainous ground like the Alps may lead to poor results.


    I'm not telling that you should keep to Level 7, not at all, but there's an optimum to be found, and my experience is the best compromise in the Alps is usually a lower resolution (typically 10-20m) than on flat ground where a 5m resolution brings great detail.


    The reason for this is simple: automated dem measurement methods generate a lot of aberrations that must be filtered out.

    Spikes amid Belgium can be filtered very efficiently. It is much more difficult in the Alps where algorithms can totally ruin a mountain shape. The finer the mesh, the more unpleasant spikes or poor renditions you may have...


    A clear example is the Matterhorn (Cervin) as featured in Intermap's 5m DEM, here seen from the Zermatt side - the famous shape is totally unrecognizable:


    Below the same point of view with another, commercial 10m resolution mesh:


    After reporting the flaw to Intermap (their NEXT 5m DEM is a very expensive commercial product) they agreed it was a shame, but correcting such issues would be too expensive compared to the few complaints they had...


    On the opposite, a 5m mesh may bring great detail to a rather flat area as illustrated below




    Cheers

    Antoine

    Config : i7 6900K - 20MB currently set at 4.00GHz, Cooling Noctua NH-U14S, Motherboard ASUS Rampage V Extreme U3.1, RAM HyperX Savage Black Edition 16GB DDR4 3000 MHz, Graphic Card Gigabyte GeForce GTX 1080 8GB, Power supply Corsair RM Series 850W, Windows 10 64 bit.

    Edited 2 times, last by Trespassers ().

  • SRTM 1" means resolution of 1 arc second, so it means ~30 meters (at the equator), not 10 meters.

    Oh, man, I'm stupid. :S:):S

    My Yosemite download from https://viewer.nationalmap.gov/basic/ was 1/3 arc-second DEM, which is 10 meter. So my Alps project came to an abrupt end, because Taburets mesh for Italy also includes the Alps with a 19 m mesh, better than 30 m.


    As counter-intuitive as it may read, a very high resolution mesh for mountainous ground like the Alps may lead to poor results.

    Good examples :thumbup:. So I will be happy with the 19 m in the alps and I love Taburets 5 m mesh for the canary islands, with which my favoured beach "Playa de papagayo" looks nearly like in real.<3