Last week I hosted a webinar on how to extend and customize ENVI using IDL (watch a recording of it here). Short summary: create and use your own software tools and algorithms for earth science and remote sensing. This was a popular topic because any project inevitably has some quirk or custom piece to it that requires special handling.
It doesn’t even matter what the project is. Trying to use historic data? Too bad they’re only on 5.25” floppy disks (remember those?). Want to add a convenient electrical outlet to your house? Too bad the builder used cutting-edge 1970s technology, aluminum wiring. That great tasting recipe? There’s always that one ingredient not found in any store. In my house, when the wheels come off of what should be a simple job, we call it an Art Project. Lots of extra custom work requiring specialized tools.
Art Projects come up all the time in the sciences. You get a new project or great idea and start doing some planning. At first, it seems straightforward. You know what you’re trying to accomplish and what kind of data you can get. That makes the workflow pretty clear, usually something along the lines of: Data ingest, calibration and geometric processing, relevant analyses, and final visualizations. It’ll be your fastest publication or product ever!
Then the details hit. Those data are all in a proprietary format from a company that doesn’t exist anymore. They’re in absorption rather than reflectance, have a pixel size in Ploughgates (remember, kids, that’s 8 Oxgangs!), and are stored on tapes that will disintegrate in the one functioning tape drive left on the continent. The algorithm you wanted to use requires a spectral band that was omitted from the one sensor that got imagery in the year you’re studying. And your IT funding was cut, so you’ll have to run it on an old Windows NT or Mac OS Housecat machine that’s lying around. Forget Pi, you want to see an irrational number? Try counting the number of Things That Can Go Wrong in a big project in the earth sciences. What happened to your straightforward workflow? It turned in to an Art Project.
For Art Projects, the key is to get or create the right tool. Like any project, use the right tool or you’ll get the wrong results, at best. If we can quickly adapt and customize, we’ll get the job done right and efficiently. If the project is going to impose onerous and rigid constraints, then we need to have flexible tools. The tool I showed in the webinar was one I developed to get a better view of Landsat data I was using. It computes every possible band ratio and then creates linked color displays based on variance (you can download it here). It was quick to develop and easy to deploy. ENVI and IDL start off as a great earth science toolbox, but when the inevitable tricky bit pops up, you can create, borrow, or buy the right tool to finish the job. It can save a lot of time, too. Maybe even enough to do a fun real art project.
I enjoyed hosting the webinar and I’d like to host more. I hope the attendees got a good start on adding to their scientific toolboxes, too. What would you do with the time saved with easy customization? What custom technology and tools have you had to create to keep moving forward in the earth sciences? Hit the comments with your stories. I’d love to hear them!