Geography Reference
In-Depth Information
of set surveying and levelling. Ground survey techniques make use of the fact that we
can measure the position of a point in three dimensions through the measurement of
angles and distances from other positions. Angles may be measured using tools such
as theodolites, while the advent of electronic distance measurers allows rapid and
highly accurate measurement of distances. h e measurement of angles and distances
may be combined using a high-precision system called a total station. Total stations
are capable of obtaining positional measurements accurate to 1 mm or so. Global posi-
tioning system (GPS) technology has a major role to play in ground survey. GPS
receivers vary markedly in their characteristics and costs. h ey range from small hand-
held systems capable of obtaining positional measurements accurate to within a few
metres to dif erential systems costing tens of thousands of pounds (or dollars) but capa-
ble of obtaining positional measurements accurate to within 1 cm. With dif erential
GPS, a stationary base station receiver is used to rei ne measurements made by one
or more roving receivers. Another important technology for generation of spatial data
is terrestrial LiDAR. h is approach can be used to generate very accurate models
of topographic surfaces and other objects on the surface of the Earth (see Pietro
et al.
(2008) for an example). Introductions to ground survey are provided by Lloyd (2004) and
Longley
et al.
(2005a) and a very detailed account is the topic by Bannister
et al.
(1998).
Sources of data error
2.9
All data are only representations of reality and are subject to a variety of factors which
may af ect their quality. h is topic outlines a wide range of methods for extracting
information from data, but for these analyses to be worthwhile the data must be of
sui cient quality. Sources of error in data include errors made during data collection,
data input errors, and inappropriate data model choices. Identifying obvious errors
caused by factors like equipment failure may ot en be straightforward, but many
major sources of error may go unnoticed. Modelling and analysis of spatial data may
introduce further errors. Manual conversion from paper-based to digital formats is
a major potential source of error. In addition, any data collection technology has a
limited precision even if that technology is employed properly and there are no exter-
nal factors (such as atmospheric conditions) that have an impact on the accuracy of
measurements. In terms of data model choice, if the spatial resolution of a raster image
is coarse in relation to the area of objects of interest then those objects will not be well
represented by the raster. A particular example is the representation of a linear feature
like a river, which may be blocky in appearance if represented by a raster.
h e accuracy of a data product can be conceived of as having two parts—bias and
precision. A biased set of measurements may consistently over-estimate or under-
estimate the 'true' value while precision refers to the repeatability of values. In other
words, if there is apparently random variation in measurements and repeated mea-
surements dif er in some inconsistent way then the measurements are of low preci-
sion. In contrast, if repeated measurements are similar then the measurements are of
Search WWH ::
Custom Search