Before we leave aerial photography, we look at how air photos can be used to make measurements of features and create map-like images of the landscape. These tasks involve fairly complex mathematics, much of which is usually done by the computer, but we will just look at the general picture and how you can use the results to make more effective use of photographs.
Photogrammetry is the science of taking measurements from air photos. Typically this involves tasks like measuring the height of buildings in a photo, or plotting elevation contours based on stereo pairs.
Remember the relief
displacement that we saw a bit earlier? This was the
fact that the heights of objects affect where they appear in an
air photo. The higher the object, the farther they get
moved away from the center of the photo. The center of the
photo is where the camera is looking at directly. Think
about it this way: imagine you are in a balloon, high above
an intersection in a city, with tall buildings on each of the
four corners. Your position is off slightly to the side of
the buildings.
You can see both the base of a building as well as its roof. The building appears to be leaning away from you. In other words, the base of the building is closer to the point just below your balloon than the roof. The roof is displaced outward from your position because it is higher than the base of the building. Notice I said outward from your position. If you were over an intersection with four tall buildings at each corner, the buildings would all be "leaning" away from you. The same thing happens to hills and other topography in a landscape recorded on an air photo. Higher objects get displaced away from the center compared to where their base would be.
All this means that an air photo cannot
be used like a map. Features in a photo are not in the same
location as they would be on a map because relief displacement
moves them from their true map (planimetric)
positions. The only exception would be for a landscape that
is extremely flat. Even a few feet of relief within a photo
displaces objects enough to create error for a detailed map.
Relief displacement creates a problem for cartographers, since we cannot map directly from an air photo. But photogrammetry turns this problem into an advantage! We can actually measure how much a feature has been displaced. And this displacement is directly related to the height of the object. So we learn not only the true, map (planimetric) position of the object, but also its elevation. So from air photos we can create both an accurate two-dimensional map and a map of topography! The topography can be expressed either as elevations of individual points, known as spot elevations, or as elevation contour lines. For nearly all USGS topographic maps, the contour lines have been created from air photos.
You might have wondered how the relief displacement is measured on an air photo. After all, it's difficult to tell that a spot on a photo is actually a hill and thereby displaced, rather than a flat area or even a depression! The answer is in two parts.
Although we have seen that stereo pairs allow us to measure features and draw maps, our original photos still aren't correct compared to the map. Should we give up on photos as an accurate portrayal of the landscape? It would be nice to be able to use the photos, for instance, to display a photo as a backdrop in our GIS and show roads and other features on top of the photo.
Fortunately, a way has been developed to correct aerial photographs so they show objects in their correct planimetric positions. An orthophotograph is a photo that has been corrected for relief displacement and other errors in the original photo. Aerial photographs usually have three types of errors:
An orthophoto corrects for these problems and can be used as a map-like image of the landscape. By the way, "ortho" is a prefix that means "true" -- think of how an orthodontist corrects problems in alignment of teeth.
Orthophotos were
originally constructed manually. The negative was projected
onto photographic paper a bit at a time, and each small area was
corrected for displacement at a time. Today almost all
orthophotos are done by computer, and are called digital
orthophotographs. A digital orthophoto is created by
first scanning the photo into the computer. Then
specialized software compares the photo to a digital elevation
model (DEM -- remember this from Exercise 5?) and a
"lens report" that tells of any errors in the camera
lens. The computer then calculates the correct image from
the scanned photo.
The USGS has a program to create digital orthophotos for all
of its topographic quads. These images are called digital
orthophoto quads (DOQs). Since the images are very large
files, they are usually divided into four pieces known as digital
orthophoto quarter-quads. You can link to the
USGS
Digital Orthophoto Program for more information on available
DOQs. Unfortunately, DOQs for Sonoma County are not yet
available. Much of the Bay Area is available,
however. Bay Area data is also available at the
USGS
Bay Area Regional Database (BARD). Most of these DOQ
data will be available fairly cheaply. They may not all be
available on-line since the volume of data is huge.
5. What happens if an object is lower than other objects in a photo? (That is, in what direction, if any, does it move relative to the center of the photo?)
6. Give an example of a natural (not human-made) landscape feature that is essentially vertical, so that you might be able to see its base and top on an aerial photograph.
Bryan Baker, Sonoma State
University, bryan.baker@sonoma.edu
Updated 17 February 1999