Moving the Arch

By Casey J. McLaughlin

Remotely sensed data is a popular background for many maps today, but it isn’t all just a pretty picture.  Images produced by satellite or airplane get processed and circulated at a very fast pace – everyone wants the latest and greatest.  Speed often means ortho-rectification isn’t a high priority.  Wikipedia gives a simple write up of the process:

An orthophoto, orthophotograph or orthoimage is an aerial photograph geometrically corrected (“orthorectified”) such that the scale is uniform: the photo has the same lack of distortion as a map. Unlike an uncorrected aerial photograph, an orthophotograph can be used to measure true distances, because it is an accurate representation of the Earth’s surface, having been adjusted for topographic relief,[1] lens distortion, and camera tilt.

Geospatial data can be a little more complicated than just having a latitude and longitude.  We model the earth in two-dimensional space (the globe is an imperfect sphere) and we’re still relying on planar map views (flat) even on our screens of choice.  Geospatial data should, therefore, contain good information about how that locational data (e.g. latitude and longitude) is collected and stored.  I’ve commented previously on the challenges of managing spatial data (datums, cell vs gps, projections, field data) and now I present, “The Case of the Moving Arch.”

A few summers ago we visited the Gateway Arch in St. Louis, Missouri.  After dipping my foot into the Mississippi, we ran around on the grounds of the park.  It’s a great monument although I confess I didn’t take the long and cramped ride to the top.  After my stroll through memory lane, I got back to thinking about imagery and plotted a reference point.  The graphic below shows a few years as seen using Google Earth’s historical imagery catalogue – as far as I can tell GE is still the best place for doing this exercise.

Nostalgia and the joy of Google Earth aside, notice how the arch “moves” in the images!  In both the 2011 images the Arch run south into the 2011 portion of my label.  August might have fewer letters than September or November, but the Arch is clearly running off the picture well to the left/east of the image!

Are you amazed yet?  I would hope not but thought I would check.  I saw the impact of this image distortion when I was reviewing some images with an inspector plotting some of his photographs onto a map using aerial imagery (the photos were geotagged with a lat/lon).   For this particular facility there were several images taken from a catwalk.  He placed the image location point onto a map (not Google in this case, FYI) and the location was definitely not on the cross-walk (more like walking on air).  Eventually, we found a satisfactory picture for the report but I’m left wondering if I had taken a photograph and not geotagged it in the field, which image could I use for adding an accurate latitude/longitude?

Normal-color kite aerial photograph of the upland study forest, Fort Leavenworth, Kansas, May 2000. Oblique view toward the south shows a fully developed forest canopy. Kansas Geological Survey Current Research in Earth Sciences, Bulletin 248, Part 1

Perhaps I’m getting into geography minutia, but there is more and more finer resolution imagery available than ever before.  Satellites and planes and kites (yes, kites) are acquiring imagery faster and finer (gigapixels!) and this trend will only continue.  In the movie “Enemy of the State” Will Smith’s character (Robert Dean) is tracked in real-time by satellite (drone maybe, but satellite?!?).  If we envision using remotely sensed imagery with such detail (seeing a dime on the street is another example) then knowing the distortion and resulting precision/accuracy seems rather important.

Casey McLaughlin is a first generation Geospatial Enthusiast who has worked with EPA since 2003 as a contractor and now as the Regional GIS Lead. He currently holds the rank of #1 GISer in EPA Region 7′s Environmental Services Division.