I need a 3D relief map made from scratch.
December 28, 2015 10:34 AM Subscribe
I would like to produce a 3D printed relief map of a mountain valley. I would like to automate this process as much as possible as I have little experience producing anything for a 3D printer.
Next month I'll be visiting a rather large farm in El Salvador. My boss has tasked me with trying to pull together a topographical map of the farm for some marketing materials. Eventually, we'd like to take this data and make a 3D printed model of the farm and surrounding land. I'm basically at a loss. If this was in the United States, I'd just hop on to the USGS train and find the highest resolution map they have and go from there, but since we're going to be in El Salvador, I'm not sure they have the same level of accessible data (If i'm wrong, please direct me; I would prefer not to reinvent the wheel here, but am willing to if need be).
We'd like this process to be repeatable, as we'll hopefully be reproducing these maps for different farms in several countries. Any gear for a project like this would preferably be low-profile, and packable. Some of the farms we'll eventually be visiting are rather remote. If I can't carry it, or it won't survive in a pelican case, it is unfortunately useless to me. I mean, in a perfect world I'd like to find an out of the box solution; I'd really rather not hack together something. If there's a drone out there that has a big red "make me a relief map" button on it, I'm pretty sure thats the solution we'd go with. Ease of use, or streamlined nature of construction would probably be favored over
Assuming cost isn't an issue, what's the best way to go about doing this?
Assuming cost is an issue, what is the best way to go about doing this?
My knowledge of mapping, 3D printing, and such topics is incredibly limited (but I have a medium-depth knowledge of graphic design…) 5-year-old-level speak might be required here.
Next month I'll be visiting a rather large farm in El Salvador. My boss has tasked me with trying to pull together a topographical map of the farm for some marketing materials. Eventually, we'd like to take this data and make a 3D printed model of the farm and surrounding land. I'm basically at a loss. If this was in the United States, I'd just hop on to the USGS train and find the highest resolution map they have and go from there, but since we're going to be in El Salvador, I'm not sure they have the same level of accessible data (If i'm wrong, please direct me; I would prefer not to reinvent the wheel here, but am willing to if need be).
We'd like this process to be repeatable, as we'll hopefully be reproducing these maps for different farms in several countries. Any gear for a project like this would preferably be low-profile, and packable. Some of the farms we'll eventually be visiting are rather remote. If I can't carry it, or it won't survive in a pelican case, it is unfortunately useless to me. I mean, in a perfect world I'd like to find an out of the box solution; I'd really rather not hack together something. If there's a drone out there that has a big red "make me a relief map" button on it, I'm pretty sure thats the solution we'd go with. Ease of use, or streamlined nature of construction would probably be favored over
Assuming cost isn't an issue, what's the best way to go about doing this?
Assuming cost is an issue, what is the best way to go about doing this?
My knowledge of mapping, 3D printing, and such topics is incredibly limited (but I have a medium-depth knowledge of graphic design…) 5-year-old-level speak might be required here.
Best answer: You are looking for Digital Elevation Models (DEM).
30M resolution is about as good at you are going to get from the global datasets. ASTER is (I think) the best option. The DEM is at the bottom of that link. Earth Explorer is also worth poking at.
LIDAR is by far the most effective way to get a truly accurate DEM, but that involves mounting lasers on a plane or a drone, flying a search pattern over an area, and then building the DEM from that data. It isn't cheap, but with drones is getting a lot cheaper. It looks like from random googling that there are definitely companies that will do a LIDAR survey for you, and may have already done it, and will sell the data for a price. I would also poke around on various government web sites- you might be able to get DEMs for free that way too.
posted by rockindata at 11:10 AM on December 28, 2015 [2 favorites]
30M resolution is about as good at you are going to get from the global datasets. ASTER is (I think) the best option. The DEM is at the bottom of that link. Earth Explorer is also worth poking at.
LIDAR is by far the most effective way to get a truly accurate DEM, but that involves mounting lasers on a plane or a drone, flying a search pattern over an area, and then building the DEM from that data. It isn't cheap, but with drones is getting a lot cheaper. It looks like from random googling that there are definitely companies that will do a LIDAR survey for you, and may have already done it, and will sell the data for a price. I would also poke around on various government web sites- you might be able to get DEMs for free that way too.
posted by rockindata at 11:10 AM on December 28, 2015 [2 favorites]
Once you have the data layers separated and you can output a barebones contour file as a vector, you can import that into your 3D program of choice and create your shape. You may wish to build one version as a series of extruded contours (think ziggurat) and another version as a mesh, which will give you a smoother contour but depending on the resolution of the imported files might obscure or elide some geographical features.
How big is the final model? Do you have to hand-carry it to the farm or can it be produced in the US, shipped to the farm, and assembled on site from parts? Is the 3D print the final or will it be used as a positive for mold making?
posted by a halcyon day at 11:31 AM on December 28, 2015
How big is the final model? Do you have to hand-carry it to the farm or can it be produced in the US, shipped to the farm, and assembled on site from parts? Is the 3D print the final or will it be used as a positive for mold making?
posted by a halcyon day at 11:31 AM on December 28, 2015
Response by poster: How big is the final model?
We'd like to attempt approximately 2'x2' model in every circumstance, but we would scale the farms to fit inside those bounds. The farm sizes can range from 50-300 acres.
Do you have to hand-carry it to the farm or can it be produced in the US, shipped to the farm, and assembled on site from parts?
The models don't need to be transported, they'll mainly be in-house marketing materials for us, and the farmers to use here in the US. With certain farms, we might gift them a 2nd model,
Is the 3D print the final or will it be used as a positive for mold making?
Honestly, either would be fine since we would only be making, at most, 2 models per farm. We were thinking that the 3D print of the landscape would provide us with the 'bones' of the model, where we'd add in the details using analog model-building techniques (ie, we would put models of the buildings on the land, add foliage, etc). We kind of just need a rough of the land.
With the above information about the 30m resolution data provided by NASA, we might be able to build the base of the model with that information, and just cover up the rough edges with something like plaster or other model-building medium?
I should have stated too that we don't need these for straight up surveying-level definition, we just need to produce a model that's fairly accurate, and representative of the land. This is a tool for people who aren't familiar with the farms to grab a sense of place and scope of the whole zone.
posted by furnace.heart at 12:02 PM on December 28, 2015
We'd like to attempt approximately 2'x2' model in every circumstance, but we would scale the farms to fit inside those bounds. The farm sizes can range from 50-300 acres.
Do you have to hand-carry it to the farm or can it be produced in the US, shipped to the farm, and assembled on site from parts?
The models don't need to be transported, they'll mainly be in-house marketing materials for us, and the farmers to use here in the US. With certain farms, we might gift them a 2nd model,
Is the 3D print the final or will it be used as a positive for mold making?
Honestly, either would be fine since we would only be making, at most, 2 models per farm. We were thinking that the 3D print of the landscape would provide us with the 'bones' of the model, where we'd add in the details using analog model-building techniques (ie, we would put models of the buildings on the land, add foliage, etc). We kind of just need a rough of the land.
With the above information about the 30m resolution data provided by NASA, we might be able to build the base of the model with that information, and just cover up the rough edges with something like plaster or other model-building medium?
I should have stated too that we don't need these for straight up surveying-level definition, we just need to produce a model that's fairly accurate, and representative of the land. This is a tool for people who aren't familiar with the farms to grab a sense of place and scope of the whole zone.
posted by furnace.heart at 12:02 PM on December 28, 2015
I'd start poking around the OpenStreetMap community - they do a lot of map data sourcing, sometimes including topo data. The OSM wiki is probably a good place to start.
posted by sibilatorix at 12:05 PM on December 28, 2015
posted by sibilatorix at 12:05 PM on December 28, 2015
Best answer: It's pretty much the last stage in your process. But I'd consider having the model milled out of something like a 15# sign foam using a CAD/CAM milling machine. This is a fairly simple application (3 axis) on a reasonable size, so I'd think you could get this solution much cheaper than a printed one.
posted by meinvt at 12:17 PM on December 28, 2015 [3 favorites]
posted by meinvt at 12:17 PM on December 28, 2015 [3 favorites]
this is a hard nut to crack.
dem/dted is the right digital format to chase. i think the ASTER data is the best bet for the rough approximation you desire.
on the other hand...for elevation there's not a very high rate of change (as opposed to man-made crap like roads and structures). so, something fairly static, and even fairly old would probably suit you. Old, static, raster...could be a paper map is starting point.
Because elevation data + foreign-place-the-CIA-used-to-be-interested-in, my first thought is the old Defense Mapping Agency (DMA) Tactical Pilotage Charts (TPC). The scale seems not very useful for farming, but maybe it is (1:500K, 500' contour interval)...the org who made these changed over the years from DMA > NIMA > NGA > FAA. NGA says look here.
this makes me want to launch a foia request, tbh ;-)
Out on a limb: Any chance you can get some NGO to fly/fund-flying the lidar specifically for your project?
posted by j_curiouser at 1:11 PM on December 28, 2015 [1 favorite]
dem/dted is the right digital format to chase. i think the ASTER data is the best bet for the rough approximation you desire.
on the other hand...for elevation there's not a very high rate of change (as opposed to man-made crap like roads and structures). so, something fairly static, and even fairly old would probably suit you. Old, static, raster...could be a paper map is starting point.
Because elevation data + foreign-place-the-CIA-used-to-be-interested-in, my first thought is the old Defense Mapping Agency (DMA) Tactical Pilotage Charts (TPC). The scale seems not very useful for farming, but maybe it is (1:500K, 500' contour interval)...the org who made these changed over the years from DMA > NIMA > NGA > FAA. NGA says look here.
this makes me want to launch a foia request, tbh ;-)
Out on a limb: Any chance you can get some NGO to fly/fund-flying the lidar specifically for your project?
posted by j_curiouser at 1:11 PM on December 28, 2015 [1 favorite]
Best answer: Terrain2STL does what you ask for. Put in coordinates on Google Maps, get a 3D printable file.
But 3D printing is not the right technology for your needs. Portable 3D printers only have a volume measured in inches, and at that scale, any normal farm is just going to be a flat square. They can't do color (except for very large expensive professional models). A small CNC mill (Shapeoko?) with slicing software (123D?) would make a reasonable-sized model out of cardboard that you could glue together in layers. Plus you could use a normal inkjet printer or even a pen to indicate features on the terrain,.
posted by miyabo at 5:15 PM on December 28, 2015 [2 favorites]
But 3D printing is not the right technology for your needs. Portable 3D printers only have a volume measured in inches, and at that scale, any normal farm is just going to be a flat square. They can't do color (except for very large expensive professional models). A small CNC mill (Shapeoko?) with slicing software (123D?) would make a reasonable-sized model out of cardboard that you could glue together in layers. Plus you could use a normal inkjet printer or even a pen to indicate features on the terrain,.
posted by miyabo at 5:15 PM on December 28, 2015 [2 favorites]
Terrainator is a different version of this same idea. It's US-only, but it's a little more polished -- if you go through all the steps it will show you a price for your model from Shapeways, a commercial 3D printing company. I got on the order of $200 for a 30-square-inch terrain model, and that's using professional equipment. A 2'x2' model would cost $4000 using the same technology. Doing it yourself would cost even more.
posted by miyabo at 5:31 PM on December 28, 2015
posted by miyabo at 5:31 PM on December 28, 2015
One thing to remember is that terrain relief seems much more pronounced when you're standing on it. As a 3D model, it might seem very flat and boring. Try multiplying the z dimension by some factor -- you can do this easily and judge the results in any hillshade tool (QGIS, GRASS, Arc, etc.) You might be concerned that the accuracy is affected by this, which it is, but if your intention is to convey a sense of the shape of the land, accuracy is secondary.
posted by klanawa at 7:09 PM on December 28, 2015
posted by klanawa at 7:09 PM on December 28, 2015
Best answer: I haven't yet accomplished what you're trying to do, but there is a tutorial here.
Complicated, but others have done it. It would be entirely possible to print in sections and glue together to get around the print volume limitation of a consumer printer. Additionally, there are services which will allow you to find local people who will print for a fee (3dhubs.com).
posted by Barry B. Palindromer at 11:59 AM on December 29, 2015
Complicated, but others have done it. It would be entirely possible to print in sections and glue together to get around the print volume limitation of a consumer printer. Additionally, there are services which will allow you to find local people who will print for a fee (3dhubs.com).
posted by Barry B. Palindromer at 11:59 AM on December 29, 2015
Best answer: I built Topophile for 3D printing terrain. It is much like several others already mentioned. It is based on 30 meter SRTM data covering North and South America, and most of Europe and Africa. I've integrated the modeling system with Shapeways for the printing of models. Their excellent SLS printers are great for terrain models.
If nothing else, playing with Topophile should give you an idea of what this resolution of data can do for you. For simplicity, I support up to 8"x8" plates, which tend to cost about $80 depending on their volume, 9 of those would give you 2'x2' for $720. Shapeways does support bigger models.
While this project is mostly on the back burner for me now, I am slowly working to add more and higher resolution data to the system. I also hope to bring full color models using satellite imagery. Dreams!
The topophile tech stack looks like:
Front-end:
- Maquest-hosted OpenStreetMaps for geo exploration (Google maps might have been a good choice, but I wanted to avoid google for this project)
- JSC3D for browser based viewing of STL models (There are more modern libraries, but I wanted to support non webgl browsers)
Server:
- web services based on cherrypy to support model requests and Shapeways integration
- *** UMN Mapserver *** serves DEM files -- Mapserver is worth the effort, it does so much for you.
- A custom python-based parametric CAD to generate STL models from elevation data. (using gdal and numpy extensively, this is the fun part, how I started playing with 3d printing models)
Please contact me thru the website if you have additional questions.
posted by concavity at 4:45 PM on December 29, 2015 [3 favorites]
If nothing else, playing with Topophile should give you an idea of what this resolution of data can do for you. For simplicity, I support up to 8"x8" plates, which tend to cost about $80 depending on their volume, 9 of those would give you 2'x2' for $720. Shapeways does support bigger models.
While this project is mostly on the back burner for me now, I am slowly working to add more and higher resolution data to the system. I also hope to bring full color models using satellite imagery. Dreams!
The topophile tech stack looks like:
Front-end:
- Maquest-hosted OpenStreetMaps for geo exploration (Google maps might have been a good choice, but I wanted to avoid google for this project)
- JSC3D for browser based viewing of STL models (There are more modern libraries, but I wanted to support non webgl browsers)
Server:
- web services based on cherrypy to support model requests and Shapeways integration
- *** UMN Mapserver *** serves DEM files -- Mapserver is worth the effort, it does so much for you.
- A custom python-based parametric CAD to generate STL models from elevation data. (using gdal and numpy extensively, this is the fun part, how I started playing with 3d printing models)
Please contact me thru the website if you have additional questions.
posted by concavity at 4:45 PM on December 29, 2015 [3 favorites]
« Older Why Don't They Make a Cheap Drip Coffemaker That... | How to deal with an opinionated, overbearing mom... Newer »
This thread is closed to new comments.
Once you have the SRTM data, you can use GDAL and/or QGISQGIS to read it, clip it, shade it, and TIFF'ify it into something you could take into Photoshop or some other graphics application. GDAL should be available on your favorite Linux distribution or Homebrew install. See here for some inspiration.
Push-button drone-based elevation data isn't quite there yet, but it is getting close. Once you have your drone-based imagery, you can use Agisoft Photoscan or Pix4d to generate a model.
posted by hobu at 11:05 AM on December 28, 2015 [2 favorites]