Jump to content

 
Photo

3d trail maps

- - - - -

  • Please log in to reply
9 replies to this topic

#1
jerry mcmanus

jerry mcmanus

    Newbie

  • Validated Member
  • Pip
  • 5 posts
  • United States

Hi Folks,

This is my first time posting, I stumbled on this site while looking for info on printing and folding poster sized images.

I've been noodling with a map project for a few years now, more or less as a hobby, and I thought people here might like to take a look. The idea of 3d trail maps is not a new one, mostly shaded relief, but I've seen the occasional oblique view as well. My project is in the latter category, not strictly a cartographic product, but computer rendered 3d views of GIS data for popular western Washington state hiking areas:

sample low res images
http://www.users.qwe...rymcm/3dtrails/

Please keep in mind that these are very low resolution samples, only 72 dpi and jpeg compressed to keep file sizes to a minimum. Also, these are NOT finished products, but very much a work in progress. The final products, if I ever get there, would be 300 dpi at a minimum and would also have many more details, such as more place names, elevation numbers, distance markers, etc.

The lighting/shading and points of view, as well as the terrain covered, are completely arbitrary. I've actually been giving a lot of thought to the final format and my current thinking is that these 12 x 18 images are probably not the way to go, for a number of reasons. I'm leaning heavily towards the format used by Trails Illustrated, nice big 24 x 36 poster sized maps that will cover an entire national park in one image. I don't have any samples of what a large format image would look like (yet), but it would cover roughly four times the area of these 12 x 18 images, so these samples are actually a good proxy for what the level of detail that would be in a 'chunk' of the larger format.

Anyway, I'd like to hear what people think, and I'll be happy to answer any questions about how they were created. Thanks!

Cheers,
Jerry

#2
Matthew Hampton

Matthew Hampton

    Hall of Fame

  • Moderator
  • PipPipPipPipPipPipPip
  • 1,306 posts
  • Gender:Male
  • Location:Portland, Oregon
  • Interests:Playing in the mountains and rivers.
  • United States

Nice maps Jerry!

My initial suggestion would be to increase the vertical exaggeration slightly. I have hiked/climbed in several of the locations you've mapped and while I really like your oblique renderings - my perception of those areas is that they are much, much more vertically rugged (Mt Adams and Rainier are massive volcanoes, and Alpine Lakes is quite rugged).

I can't recall any references at the moment - but the gist of it is: we pedestrian humans are very tiny compared to mountains. Most of our perceptions of mountains comes from looking up at them from the surface of the earth, thus it is common to exaggerate the relief to account for our perception. Although if your target audience was "in-flight magazines," they would be well matched.

Is that an NSD render?

co-cartographic creator of boringmaps.com


#3
frax

frax

    Hall of Fame

  • Associate Admin
  • PipPipPipPipPipPipPip
  • 2,304 posts
  • Gender:Male
  • Location:Stockholm, Sweden
  • Interests:music, hiking, friends, nature, photography, traveling. and maps!
  • Sweden

The look gorgeous, congratulations!

For out on the trial - I wonder though what additional value an oblique view offers, maybe an isometric view would be better, or just an ordinary map.

For the wall, or for signs, I think this would work very well.
Hugo Ahlenius
Nordpil - custom maps and GIS
http://nordpil.com/
Twitter

#4
jerry mcmanus

jerry mcmanus

    Newbie

  • Validated Member
  • Pip
  • 5 posts
  • United States

Nice maps Jerry!

My initial suggestion would be to increase the vertical exaggeration slightly. I have hiked/climbed in several of the locations you've mapped and while I really like your oblique renderings - my perception of those areas is that they are much, much more vertically rugged (Mt Adams and Rainier are massive volcanoes, and Alpine Lakes is quite rugged).

Is that an NSD render?


Thanks Matthew, I'm glad you like them.

Thanks for the tip, I've also felt the real elevations seem too 'flat', but there's an important trade-off to consider: The taller the terrain features are, the more likely they are to conceal the features behind them. This can be compensated for somewhat by increasing the angle of the camera to the ground plane, but then you are almost looking down on the terrain and the 3d effect becomes little better than shaded relief.

The 3d software I use to render the scenes is an Open Source program called Blender:

http://www.blender.org/

It has all the features I needed and you sure can't beat the price.

Cheers,
Jerry

#5
jerry mcmanus

jerry mcmanus

    Newbie

  • Validated Member
  • Pip
  • 5 posts
  • United States

The look gorgeous, congratulations!

For out on the trial - I wonder though what additional value an oblique view offers, maybe an isometric view would be better, or just an ordinary map.

For the wall, or for signs, I think this would work very well.


Thanks Frax,

You have a good point, these 'maps' are not really maps at all, at least not strictly in the cartographic meaning of the word, they are really only illustrations, albeit ones that use actual GIS data.

I've been struggling with the question: are these trail guides or wall posters? If they are trail guides they would need to be folded into the standard 4 x 9 format with lots of trail info printed on the back side, and I don't really have the resources to compile that info, besides which there are already many good Wash. State trail guide books on the market. This is one of the main reasons I'm leaning towards the wall poster format, the back side can be left blank and people can decide for themselves if it is worth it to fold it up and take it on the trail or not.

Very early in the project I contacted the local Mountaineers Club which publishes a series of popular trail guide books, and they were enthusiastic about my maps, but they could only pay $50 per image and the final format would have been about the size of a postage stamp, so I had to say no thanks.

Cheers,
Jerry

#6
Mike Boruta

Mike Boruta

    Key Contributor

  • Validated Member
  • PipPipPip
  • 99 posts
  • Gender:Male
  • Location:Ouray, Colorado
  • United States

I think the maps look really good! Whether or not you call them maps or illustrations, I love how these oblique-perspective shaded relief maps help me visualize the terrain.

I've been struggling trying to figure out how to produce this sort of map. I recently tried using ArcScene to make an large scale oblique-angle map but I was not satisfied with the results or the software. I just installed Blender on my computer and it is very very foreign to me. My question is, since this software is not really designed for cartographers, how did you go about getting GIS elevation data into Blender? Also, did you add the contours and other map data into Blender as well? Besides Blender, what other software did you use in your workflow?
Any tips would be much appreciated. And you're right, you can't beat the price!

#7
jerry mcmanus

jerry mcmanus

    Newbie

  • Validated Member
  • Pip
  • 5 posts
  • United States

I think the maps look really good! Whether or not you call them maps or illustrations, I love how these oblique-perspective shaded relief maps help me visualize the terrain.

I've been struggling trying to figure out how to produce this sort of map. I recently tried using ArcScene to make an large scale oblique-angle map but I was not satisfied with the results or the software. I just installed Blender on my computer and it is very very foreign to me. My question is, since this software is not really designed for cartographers, how did you go about getting GIS elevation data into Blender? Also, did you add the contours and other map data into Blender as well? Besides Blender, what other software did you use in your workflow?
Any tips would be much appreciated. And you're right, you can't beat the price!


Thanks!

OK, since you asked, :D here's the long version:

I have a background in interactive multimedia, things like encyclopedias on CD-ROM and touch-screen kiosks, so it was crucial when I realized right from the start that this was going to be mostly a graphic design project that happened to use GIS data, as opposed to a GIS project that just happens to produce a raster image at the end.

That freed me to avoid what would have been a very steep and expensive learning curve trying to use pro level GIS software and to focus instead on using the tools I was already familiar with, such as 3d rendering and image editing software, and also to focus on learning only what I absolutely needed to know about the raw GIS data itself.

My background also gave me the skills I needed to program some of my own custom tools which I used to convert the GIS data into the raw materials that I needed to build the graphics. So, in thinking about the workflow, I essentially started with what would eventually be the end product and worked my way backwards which, oddly enough, resulted in no GIS software being used in the production of my maps.

I am not a professional software programmer by trade, so I relied on the tool I was most familiar with which is a multimedia authoring program called Director from Adobe to build my custom tools. It has just barely enough capability with its built-in scripting language (called Lingo) that I was able to import images and text (ascii) based data sets and do basic processing on them.

The data I used for the terrain is the old 24k 10 meter DEM's which I found available for free at a University of Washington website, a site that also happened to have the corresponding 24k DRG's which made convenient one-stop shopping for western Wash. State data:

http://gis.ess.washington.edu/data/

To create the terrain models I wrote a custom tool that converts the DEM data into XYZ triplets that can be imported into the 3d rendering software. The triplets by themselves aren't enough, you also need to build the three sided faces (polygons), so my custom tool actually writes everything to a file using the old VRML (virtual reality modeling language) format that most 3d programs can import. I think Blender actually has a plug-in that imports DEM's directly, but I don't use it because I felt I needed more control over the process.

Once the terrain model is created and imported into the 3d software you need to assign a "texture" to it whcih determines the colors on the suface of the model. This is where the DRG's come in, and again I wrote a custom tool that crunches an entire DRG, one pixel at a time, and essentially strips everything out except the colors for ground cover and water features, including ice. These colors layers are saved as images and then imported into the image editing software Photoshop (also from Adobe), and they form the basis for the "texture map" that is imported into Blender in PNG format and projected onto the 3d terrain model.

The third step is to create the contour lines, and because I wanted maximum control over this process as well, I wrote yet another custom tool that crunches the DEM data and exports what is essentially a color coded heightmap, exactly the same size as the DRG, with different bands of color at specified elevation intervals. I then import this image as another layer into my Photoshop document and brush stroke the edges to get the desired color and thickness of the contour lines.

The fourth step is to add roads and trails, and for this I use either 24k or 100k DLG data, whichever is available. And yes, I wrote yet another custom tool to convert the vector data into lines drawn on an image which is also imported into Photoshop and processed as another layer in the texture map.

Once these steps have been completed for a number of 24k topo quads I can start to build the final scene in Blender. The terrain models have an enourmous number of polygons (over 250K), and I don't have a supercomputer, so I use very low-res proxy models to assemble a number of topo quads into one 3d scene and to tweak the position of the virtual lights and cameras without crashing my computer. I then render each full-res topo quad out to an image one at a time and individually assemble those pieces manually in a new Photoshop document.

From there on out it's all Photoshop, adding dozens of layers for text labels, icons, titles, legends, compass rose, etc., etc. The final product is a giant 300dpi raster image, comnplete with embedded fonts for the vector text layers, all wrapped up in a PDF format file that I hope can go straight to film when the time comes to send it to the printer.

Whew! Sorry for the long-winded post. As you can imagine this project has been a lot of work, all of these steps repeated for over 200 quads (and counting), but the results have been promising enough so far to keep me going.

BTW, if you are not familiar with 3d software then Blender is probably not the best place to start. Like other OS software it has a non-standard interface which can be extremely frustrating to use. I recently stumbled on the nifty little 3d program called SketchUp which can be downloaded for free from Google:

http://sketchup.google.com/

It has the added bonus of exporting directly to Google Earth:

http://earth.google.com/

which makes for some interesting visualization projects, such as this lark I did for my hometown newspaper:

http://blog.seattlet...c_eco_dome.html

Cheers,
Jerry

#8
Mike Boruta

Mike Boruta

    Key Contributor

  • Validated Member
  • PipPipPip
  • 99 posts
  • Gender:Male
  • Location:Ouray, Colorado
  • United States

Wow, that is a very nicely detailed description of your workflow. For the non-programmer type, is a little disheartening to learn that it involved so many custom made applications, but at the same time I admire the approach. I can see how some of these steps can also be accomplished in ArcGIS, which I have access to through school. This is a lot of food for thought, thanks again!

#9
Unit Seven

Unit Seven

    Legendary Contributor

  • Moderator
  • PipPipPipPipPip
  • 266 posts
  • Gender:Male
  • Location:New Zealand
  • New Zealand

Great repy Jerry,

One suggestion I might make for you to think about would be rather than adding those dozens of layers for icons and labels in Photoshop woul be to think about placing the raster as a linked image in Illustrator and then adding the vector items in there and exporting to pdf from there with vectors intact.

This will let you use Illustrator symbols and text styles etc which should save some time and make the map more reusable. Will also allow thes text to be remdered at the devices output resolution and give better control for trapping etc at the rip.

That aside the process and the maps look good—yes I still think they are maps albiet not really for ue on the trail but on the wall and trip planning they are a great tool.

Cheers,

Sam.
S a m B r o w n

U N I T S E V E N
unit.seven@gmail.com

Miramar, Wellington
N E W Z E A L A N D

#10
jerry mcmanus

jerry mcmanus

    Newbie

  • Validated Member
  • Pip
  • 5 posts
  • United States

Great repy Jerry,

One suggestion I might make for you to think about would be rather than adding those dozens of layers for icons and labels in Photoshop woul be to think about placing the raster as a linked image in Illustrator and then adding the vector items in there and exporting to pdf from there with vectors intact.

This will let you use Illustrator symbols and text styles etc which should save some time and make the map more reusable. Will also allow thes text to be remdered at the devices output resolution and give better control for trapping etc at the rip.

That aside the process and the maps look good—yes I still think they are maps albiet not really for ue on the trail but on the wall and trip planning they are a great tool.

Cheers,

Sam.


Thanks for the suggestion Sam, but unfortunately I don't have Illustrator, or any experience using it.

No worries, I think Photoshop has most of the features you mentioned, the layers can be grouped into folders and copied between files for easy reusability, and the text is not rasterized but is preserved as vector data in the pdf for scalability. I've verified this by printing full res uncompressed proofs on a high end commercial laser printer and got spectacular results, so I'm actually quite happy with the arrangement so far. If the text looks blurry in the samples, as I mentioned before, it's only because I deliberately downsampled, rasterized, and jpeg compressed all the sample pdf's for minimal file size.

If I start adding little icons I might wish for some Illustrator-like features, but even they would probably be small enough that at 300dpi the difference would be hardly visible in the full res images.

Cheers,
Jerry




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

-->