Geography Reference
In-Depth Information
The initial digitising procedure was undertaken using a HP 5100c flatbed
scanner with a default scan speed for high quality scans. Resolution was set at
a 150 dots per inch (DPI). Even with the larger scale of the survey maps, each
map required at least four and sometimes six A4 scanned sheets that needed to
be geo-located and joined in the GIS software. This also became a potential
source of accumulative errors.
In an attempt to overcome these potential errors, digital photography was
used on the original survey maps. The camera used in this process was a
Digital SLR Nikon D1X with a 28-70mm 2.8 lens. The picture quality setting
for this process was the RAW format with a resolution of 300dpi. The
photographs were taken outside in a shaded area with natural light and a
distance of 1 to 1.5m between the map and the camera. Generally, shutter
speed was 100 th of a second, but the aperture was varied between f18 and f11
with a focal length between 42 and 31 mm, which when translated into 35 mm
film equivalents, was between 63 mm and 46 mm. The advantage of speed and
efficiency that was anticipated by this procedure was negated by a lack of map
clarity and consequent limitations on interpretation of the final photographic
images. This was in spite of enhancement using Adobe Photoshop 7 filters and
image adjustments to the image sharpness, brightness and contrast.
With these disappointing results we returned to the initial process of
joining A4 scanned images, see also (Hazell and Brodie, 2012).
To keep within a practical research time frame we adopted a multi-layered
approach by capturing thematic layers from scanned 1:250,000 survey maps.
In later analysis, maps with a scale of 1:50,000 were used to provide greater
detail in specific areas of interest. All maps were scanned and enhanced using
Adobe Photoshop 7 to improve legibility and associated accuracy during final
digitisation. Nevertheless our interpretation of specific features such as
contour position and value imposed limitations on final accuracy. Points on
contour lines were inserted at changes in direction and intermediate points
were included as frequently as possible to minimize error (Douglas and
Peucker, 1973). The individual maps were joined into a matrix, as shown in
Figure 3, from which an initial analysis could be undertaken. The contour data
was then used to create a surface model of the landscape as a Triangular
Irregular Network (TIN) that was formed from our digitised contour data
(Figure 4).
Search WWH ::




Custom Search