UNDERWATERSURVEY 131
measurement used as a check. It is possible to plot the four
points at the same height by hand and to determine how
different the check measurement is from the expected value;
this value is called the residual. If the points are at dif-
ferent heights then the depths can be measured and the
measurements can be corrected for the difference using
Pythagoras’ theorem and the points can be plotted.
When there are more than four points, or more than
one check measurement is made, then a problem occurs.
The points can be plotted in a number of different places
depending on which measurements are used. There are
no rules to say which measurements should be used, so
statistics are employed. A mathematical technique called
‘least squares’ can be used to compute the best answer
for any given set of measurements. This is known as
‘adjustment’. The technique is mathematically intensive and
is best done using a dedicated computer program. Most
sites using 3-D trilateration utilize the ‘Site Recorder’
or ‘Site Surveyor’ software from 3H Consulting Ltd or
‘Web for Windows’ (DSM), although other programs
can also be used.
The computer program takes in initial guesses or estim-
ates for the positions of the survey points, along with
measurements between the points. From this, the software
calculates new positions for the points on the site along
with some position-accuracy information for each
point. Based on the new positions for the points, the
programme can also give an idea of the quality of the
measurements or how well the measurements fit together.
Because many sites are essentially flat, the distance
measurements tell us very little about the depths of each
point. This is why it is essential to measure the depth of
each control point and detail point, and to include these
measurements when processing. When the calculated
position accuracy for each of the control points is within
the accuracy specified in the project design, this phase
of the work is complete. At this stage it is essential to ‘fix’
the positions of all of the control points in the computer
program so that the addition of further points and meas-
urements does not affect the carefully calculated control
point positions.
With the control points carefully positioned, they can
be used to position detail points on the site. The detail
points are added to features, artefacts and structure, so,
by positioning the detail points, the positions of the
features themselves can be calculated. To position detail
points, a direct distance measurement is needed from
each detail point to the four nearest control points.
Ideally the measurements should be made from points
all around the detail point rather than from one side
only (figure 14.18).
Distance measurements between detail points on the
same object can be used, such as the distance between the
detail points at each end of a cannon. However, measur-
ing distances between detail points on different objects
should be avoided. This is because the objects themselves
may be moved as the site is worked on and it also makes
processing much more difficult because it is harder to find
incorrect or ‘blunder’ measurements. Mistakes in meas-
urements between detail points will affect the position
of both of the points, so one incorrect measurement can
alter the positions of a whole chain of detail points linked
together by measurements. A blunder measurement be-
tween a detail point and a ‘fixed’ control point will only
move the position of the detail point, constraining the effect
of the blunder measurement to one detail point only. As
with control points, it is essential that a depth measure-
ment be made at each detail point.
The computer program used for processing or ‘adjust-
ing’ the measurements will calculate the position of eachLong ‘Pointy’DisconnectedFigure 14.17 Poor control-point network shapes (3-D survey).