- 972--
Data Extraction and Recording
Using a Carl Zeiss Interpretoscope the black and white
.aerial photographs were stereoscopically examined to
identify those units of open space as recorded in the
notation (Figure 2).
A transparent grid overlay was placed over the photographs
so that the data could be recorded on a grid square basis.
The grid represented a 20m x 20m square at ground scale, and
that use which fell under the cross at the centre of the grid
was the use recorded for the entire cell.
Figure 4 illustrates the procedures followed in recording
the data. Where the cross landed cleanly in a particular use
then that use was the recorded one. Where the cross landed
on a boundary the decision followed a set pattern, as can be
seen in Figure 4.
The location of each grid cell was recorded on the coding form
so that spatial information could be retrieved for mapping
purposes.
The colour infra red photography of the three enumeration
districts were also stereoscopically examined and similar
information was derived from them and entered on the coding
forms.
The two sets of survey information - one from the black and
white photography, the other from the colour infra red
photography - were then fed into a computer data bank.
Computer programmes were developed to provide a variety of
statistical data from the information fed into the system.
Some consideration was given as to whether to use a grid square
or a single point (in this case a cross +) as a basis for recording
the information.
In the former case the 'dominant use! of the grid square would be
recorded, whereas in the latter case the use directly under the
cross would be recorded. It was found that the latter was very
much quicker and, unlike the former did not involve any judgement
decisions or estimates of areas.
Some random tests showed that there was no advantage in accuracy
in using the 'dominant use! of the grid square. An important
factor was the density of the observations - in all the work
carried out in this study a density of 625 points per Km^ was
used to record the data.
At a subsequent stage a programme was developed to carry out a
large number of random samples, from 625 to 10 points per Km2,
and the results of this are now being analysed.
The aim of this particular part of the investigation is to give
the planning authority some idea of the relative accuracies
attainable when using grids of varying density.
^ Polo uh 2 21 AM un Re
NA A UN
& Pr 0M rS
Ar or Fr
tn
Py
JS Gd. wu Put
JS uu af on M—M FP. 7D
An
«* ¢n
10