Most of us have been exposed to geographic relief maps during our grade school years. Remember those cheap vacuum-thermoformed plastic 3D maps that used to hang on our classroom walls? They usually depicted our entire state, and invariably some curious kid over the years poked his finger into it, leaving a permanent dent and/or crack for future students to wonder about.

Classic 3D relief maps are still being produced and they do serve a useful purpose, but... in addition to being susceptible to deformation, they are also quite limited in terms of the detail that they can show and the smallness of the area that they can cover. A few years ago, I bought one for my son that covered the entire San Francisco Bay Area all the way to Stockton (almost 6,000 square miles), which is about as small of a region as relief map manufacturers are willing to produce for the mass market.

But suppose you wanted a relief map of just your neighborhood... you know, something closer to 6 square miles rather than 6,000. How would you get that produced? That question sent me on a two-year-long quest that resulted in this "data-driven" art project. Now, thanks to instructables, you can do the same in two weeks!

Let's get started...

Step 1: Inspiration

If you're like me, you love maps---maps of any type: ancient maps, modern open source maps, globes, atlases, terrain models, Google Maps, Bing Maps... and, especially, Google Earth. "Geo-imaging" provides an opportunity to inspect--and marvel at--our planet from a perspective that we don't normally get unless we hike to the top of a mountain or grab a window seat on a plane (two things I try to do as often as possible).

One evening, while drinking a glass of wine and flying over my topographically gorgeous neighborhood (on Google Earth, as shown here), I found myself wondering if there was a way to translate my online experience into something I could hang on my wall.

It didn't take too long to find out that the U.S. Geological Survey (USGS) has an enormous wealth of public domain imagery and data available online (http://viewer.nationalmap.gov/viewer/). Happily, my neighborhood falls into the USGS "Urban Areas" dataset, a subset of the National Map that includes the highest resolution aerial imagery and elevation data available in the public domain. When I discovered this, my ambitions regarding model size expanded.

Step 2: Aerial Imagery

Once you've committed to the idea of creating your own neighborhood terrain model, the next step is to define the precise extent of the area you're interested in displaying... and find the highest quality "orthoimagery" of this area that you can. Per the National Map website:

Orthoimagery data typically are high resolution aerial images that combine the visual attributes of an aerial photograph with the spatial accuracy and reliability of a planimetric map. USGS digital orthoimage resolution may vary from 6 inches to 1 meter. The National Map offers free downloads of public domain, 1-meter orthoimagery for the conterminous United States with many urban areas and other locations at 2-foot or finer resolution.

Visit this link for further information and data download:
http://edc.usgs.gov/#Find_Data/Products_and_Data_Available/High_Resolution_Orthoimagery

Note: There are several private sources where you can buy more recent or higher resolution orthoimagery. Simply read the copyright line at the bottom of the page when you're viewing your neighborhood in Google Maps to find out who provides this. That said, if you can find something satisfactory from the USGS, go for it... free is a hard price to beat! Don't worry... you'll get your chance to spend money later when it's time to produce your model.

Here is the orthoimagery I downloaded from the USGS for my model. It was available at .25 meters per pixel resolution... which is great!

Step 3: Elevation Data

The next step is accessing elevation data of your desired area. This too is available publicly at varying resolutions. If possible, try to get 1/3 arc second resolution (about 10 meters). A few areas even go as fine as 1/9 arc second (about 3 meters), although I can tell you that 1/3 arc second is adequate. That's what I downloaded for my model and it holds so much detail that you can actually feel the jetty in my nearby harbor with your finger.

Learn more about the National Elevation Dataset here:
http://ned.usgs.gov/

FYI: In case you didn't know, our planet is quite smooth; I've heard that, relatively speaking, it's smoother than a ball bearing. For that reason, typical 3D relief maps of large areas are given a considerable vertical exaggeration (3x to 15x depending on scale) to make them look interesting. Obviously, the smaller the area, the less vertical exaggeration you need. My initial plan was to use no vertical exaggeration at all and stay true to the topography, but after extensively previewing this small region on Google Earth, I felt it needed a small push. In the end I settled on a 15% vertical exaggeration (1.15x), far less than typically used in terrain models.

Step 4: Cutting

Fast-forward now through close to two years of intermittent research, which included attending specialty printing tradeshows, visiting 3D printing service bureaus, etc. Let me save you the trouble: Get in touch with the folks at Solid Terrain Modeling Inc. (http://www.solidterrainmodeling.com/index.htm)

The founders of this specialty shop based in southern California invented a unique 3-axis ink-jet printing head that can be programmed to "fly" the same path as the cutting tool used to cut the base material: high-density polyurethane foam (similar to the stuff used in surfboards).

Their process is essentially this:

Rough cutting: An initial pass of a coarse cutting tool that roughly approximates your elevation data.
Fine cutting: A second pass that brings out the full detail of your elevation data.
Priming: Coating the entire block of foam in a special white ink-receptive paint.
Ink-jet printing: Full-color high-resolution printing.

Here is a picture of my model pre-priming. FYI: The model is roughly 25" wide by 40" long.

Step 5: Printing

Sounds simple so far, right? Ah, but the devil is in the details.

Here's a picture of my completed model... which took three attempts! Fortunately, I only had to pay once.

I was told after the fact, that on the first attempt the project manager was not happy with the color. On the second attempt, they experienced an ink-jet head crash. This is not that unusual. Apparently the tolerances are so tight--the ink-jet head needs to be so close to the surface--that occasionally it slams into the side of a hill. When that happens the operator must tweak the settings in their machine control software and hope for better luck next time.

At this point, in order to make a third printing attempt, they had to first re-cut and re-prime the entire model (they used the same block of foam). Once completed, they sent me this picture. Needless to say, I was absolutely thrilled and couldn't wait to take delivery!

Step 6: Finishing Touch

Just to be clear, in case it's not apparent from my pictures, this model is composed of two elements only: aerial photography and land elevation data. There are no three-dimensional buildings, trees, bridges, docks, boats, cars, etc. Even without such accoutrements, it still presents a very convincing facsimile of my beloved neighborhood.

However... there is one especially prominent landmark that would be familiar to anyone who has visited Half Moon Bay in the last five years or so: the big white radar "ball" atop Pillar Point. Naturally, I had to have that on my model. So I carefully studied its size and position above the landscape and then took a trip to the local fabric store to find just the right ball pin to use. You can compare the result of my efforts to the real thing in the images here.

Step 7: Hanging

Here is my completed model as it hangs in the wall of our entryway. As I mentioned at the start of this instructable, I call this a "data-driven" art project; I didn't carve it or paint it, but I did conceive the project, choose the extents, source the data, research production methods, choose a vendor... and pay handsomely for it.

The total cost of this project when I had it produced in late 2006 was $4,300 (including tax). Because of its large size, it would have required an extra $700 to build a custom crate and ship this to my house. I chose instead to make a road trip to L.A. to visit some friends and pick this up on the way home.

I should mention that this is also "functional" art in that it acts as a map (of course). I use it to show our visitors how to get out of our crazy neighborhood, I point out our favorite restaurants, I show them how to get to the beach trail, where to ride their mountain bikes, where to rent kayaks, etc. Very handy!