11.8.2
Individually solving values
No matter what the solving order may be, each value should be computed independently. All
tests
should be based on the mean of the previous test(s).
It is important to understand why a certain solving order is used in all survey software. Each
computation is based on the previous test result. This is the reason that latency is computed
before
pitch and roll is computed before heading; the primary test (latency or roll) has a large
influence on
the result for the secondary test (pitch and heading). The roll computation can
also have an
influence on the pitch computation, primarily if the position of the sonar head, of
the reciprocal
runs, was not coincident. The heading offset will also have an influence on the
pitch computation
for the same reason.
Generally, multibeam surveys are conducted with very accurate time synchronisation using GPS
time
and the Pulse Per Second. In this case, the latency test is used to prove the lack of latency
or that is
sufficiently small enough so as to be of no consequence. Using accurate timing, it is
not necessary to
collect more than two latency collections. This paper will concentrate on the
angular offset
computations. However, if accurate timing is not used there should be the same
number of
collections as with the other tests.
With a good number of individual tests, solve for one computation (i.e. only roll) and derive a
mean
and standard deviation for that one test. Determine if the standard deviation is within
acceptable
accuracy requirements, then use that derived mean to solve for the next
computation (i.e. pitch). As
an example, using the results on page 7, the first step would be to
solve for Roll first, derive a Roll
mean and then use that mean in all of the Pitch computations.
Find the mean and standard
deviation
for Pitch. Use the mean Roll and Pitch values to
determine the Heading offset.
In the above example, the roll mean, of the five tests, is 0.74°, with a standard deviation (δ) of
0.01°.
The roll mean would now be used when determining the value for pitch. Use the roll
mean and
solve all of the pitch computations; the pitch mean is -0.91° (excluding the out of
norm value), δ =
0.13°. The roll and pitch computed means are now used to solve for the
heading offset. The solved
heading offset is 0.92°, δ = 0.08°.
If the heading offset had been 1.5° or greater, it would be advisable to re-compute the pitch
offset,
using the computed heading offset value. This is due, again, to the fact that if the sonar
head did not
track the same exact position in the reciprocal runs, the heading offset will have
an influence on the
pitch offset result.
11.9 Truthing the patch test
After deriving the values for roll, pitch and yaw, the values should be entered into the
appropriate
areas in the data collection software. Ideally, find a singular object that can be
boxed in (running
data collection lines, on all sides of the object) and process the data. The
object depiction, with all
survey lines, should not vary from the object depiction from any one
line.
Page 140 of 210
Version 5.0 Rev r002
Date 05-08-2014
Part No. 96000001
To Next Page
Loading...
Need help?
General
