
Photogrammetry is the art science and technology of taking measurements
off photographs. The techniques are based on the geometry of perspective
scenes and on the principles of stereovision, and actually pre-date
the invention of photography. There are two kinds of photographs
used in photogrammetry, aerial and terrestrial. Aerial photographs are
usually acquired from aircraft but can
also come from satellites, hot air balloons or even kites. Terrestrial
photographs come from cameras based on the ground, and generally
are used in different applications from aerial. There are two
main data extraction methods used for analysing these photographs:
a. Quantitative: that is size, length, shape, height, area, etc.
b. Qualitative: geology, vegetation, drainage, land use, etc.
This chapter is primarily concerned with the quantitative evaluation
of survey photographs.
Photographs taken with a survey camera differ to those taken with amateur (or
'non-metric') cameras
in that they have a rigid known geometry that allows measurements
of a predictable accuracy to be made.
![]() |
![]() |
The rays of light pass through one central point in the lens of
a survey camera, and the distance between this point and the film
is calibrated to decimals of a millimetre. This distance is known
as the focal length. The film plane is as flat as possible, often
using flattening devices to achieve this. The camera body also
carries reference marks known as fiducial marks that define a
coordinate mesurement axis and allow film stretch to be determined.
The camera is also very large, the size of the negatives is 23cm
by 23cm, which is somewhat bigger than the average snapshot. The
cameras are constructed in this manner so that measurements of
a high accuracy and precision can be obtained.
The average scale of the aerial photograph can be computed either
by taking the ratio of the flying height above the ground and
the focal length (f/H), or by taking the ratio of a known distance
in the photograph and the distance on the ground.

As the flying height above the ground is not usually known accurately,
the second method is employed where a more reliable scale is needed.
Neither of the two scale calculation methods give an accurate scale as there are
two distortions that affect measurements made
on a single aerial photograph.

a. Height distortion:
Because a lens and a photograph give a perspective or central
projection, objects that are above or below the plane will be
shifted by an amount approximately
Dr =
b. Tilt distortion
Although a lot of care is taken in the flying of aerial photography
the photographs are rarely taken exactly horizontal. The tilts
that occur in the aircraft, although kept to a minimum by the
levelling of the camera system, do affect the position of objects
on the photograph. This is most apparent with panoramic photographs
that show parallel roads converging to a point on the horizon,
this affect still appears in photographs that have small tilts
but occur to a much smaller amount. The correction is approximately
Dr =
However as the tilts are generally unknown this correction is
difficult to apply.
These distortions have been mentioned to indicate that an aerial photograph is rarely true to scale so cannot often be used as one would use a map. There are many situations however, especially in forestry or geography, where these distortions are minimal when compared to the accuracy of the data required, and can sometimes be ignored.
There are other distortions such as film shrinkage, earth curvature,
refraction effects and so on but these are only significant when
dealing with precise photogrammetry.
Aerial photographs are normally flown in runs with a pre-determined
amount of forward and side overlap.
This forward overlap, usually 60%, enables the aerial photographs
to be viewed in 'stereo', or three dimensionally. The accurate
geometry of the photography also enables three dimensional coordinate
data to be extracted; the same height distortions that affect
the single photograph can now be used to determine the height
of the object in a stereo pair.
The detailed geometry of a pair of photographs will not be described
here, however there is an approximatel method of extracting heights
off a stereo pair which is of use.
The difference in position of an object with height displacement
on two photographs is known as parallax. If this parallax can
be measured the height of the object can be determined if the
scale of the photograph is accurately known, the air base is horizontal
and the tilts are minimal. The derivation of this formula is shown
in most photogrammetric textbooks.
z = H where b' is the photo base, p' is the difference in parallax
An instrument known as the parallax bar is used to measure the
parallax - p'.

Topographic mapping of scales of 1:5000 and smaller is these days
performed almost entirely using aerial photogrammetry. Overlapping
photographs are acquired of the area of interest (this is done
on a routine statewide basis by the Department of Property and
Services), and certain points visible in the pair of photographs
are coordinated to act ac scale and tilt c These points are used
in the plotting instrument to scale and orient the photographs,
to enable the operator to trace the necessary detail from the
'stereo-model'. A good machine operator is capable of providing
line work of cultural features like roads, buildings and so on,
as well as contours showing height. The production of contours
relys on the operator's ability to perceive depth in the pair
of photographs (it all happens in the operator's head). Each major
theme, or printing ink colour, is usually plotted either on separate
sheets or in differing colours to enable similar features to be
distinguished. Standard series maps used certain ink colours to
depict certain types of features, like blue for water and green
for vegetation. Maps also show the standard plan features like
north orientation and a scale, but also show coordinate reference
systems like AMG or latitude and longitude.
It is possible to optically remove tilt distortion from photographs
of flat terrain by reprojecting the negative onto a tilted base
board. This process is known as rectification (making right).
A rectified photograph is true to scale only where the terrain
is flat and even. These photographs can be used for planning works
or projects, and can be annotated similarly to maps with borders
and legends to become Photomaps.

Recent developments in photogrammetry have seen the development
of instruments that can eliminate height displacement as well
as tilt distortion from a photograph. The process is involved
and consists of rectifying small strips of the photograph at a
time and reconstituting this into a new photograph. The resultant
image is true to scale, and is often annotated with contours,
coordinate grids an a legend, and published as a standard series
map.
The last decade or so has seen many advances made in the technologies
associated with photogrammetry, and the microchip has had a major
impact. It is now possible to achieve similar precision and accuracy
of measurement using conventiomnal cameras and computer compensation
for lens and film effects as was previously possible with special
survey cameras and restitution equipment. This has enabled the
application of photogrammetry in biometrics, architecture, geology,
archaeology, mechanical and civil engineering, nuclear physics,
zoology and other scientific disciplines where remote measurement
is an advantage. Photogrammetry is now seen as another measurement
tool available to the measurement scientist not just a way of
making maps.
|
|
|