International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXX V, Part B2. Istanbul 2004
however, the imagery was to be used for a number of different
processes, the economies of scale may be enough to j ustity its
use.
3.2 Map Currency Audit
Each year. the UK Government defines a set of “Agency
Performance Monitors", by which to measure the performance
of Ordnance Survey. These performance targets include the
following:
* to ensure that a minimum of 99.6% of significant
real-world features (Category A) are represented in
the database within six months of their completion;
* ioensure that there is an average of no more than 0.6
standard units of un-surveyed major change over 6
months old, per standard map unit.
Note that the standard unit of change is the *house unit”, which
traditionally represented the amount of change observed on the
building of a new house. The unit has now been formalized to
encompass many other types of change. Some examples of
house unit values arc:
* New houses and associated features, including
boundary features, name/number and associated
garages = 1.25 units per house.
+ New commercial, industrial, public sector and farm
developments with buildings and associated features
= 20 units per ha
e New single carriageway roads, railways ( per pair of
tracks) and canals, including associated paths, fences
and boundary features = 5 units per 100 m
Also note that the standard map unit depends on the scale at
which the data are captured. The unit equals 25 square km in
mountain and moorland (1:10 000 scale), one square km in
rural (1:2500 scale), and 0.25 square km in urban (1:1250
scale) areas.
A Quality Assurance (QA) team within Ordnance Survey is
responsible for monitoring these values and ensuring that the
Agency is meeting the requirements. In order to do this, a
sample of 4000 map units are randomly selected every six
months. The areas covered by these maps are then visited by
field surveyors in the QA team and examined for any features
more than 6 months old, which are not on the map.
Panchromatic and pan-sharpened QuickBird images of
Cambridge (rural) and Manchester (urban) were used to test
whether the quality monitoring process could be successfully
augmented using satellite sensor imagery. The images were
examined and compared with the current large scale mapping
data. Any features identified in the image which were not on
the map were recorded. The results were then compared with
quality audits performed using traditional field verification
techniques.
3.2.1 Results of Map Currency Audit
The QA team found that QuickBird imagery could be used to
identify changes, and thus aid in the assessment of the map
currency. In several cases changes were detected in the image
which were not detected on the ground; and in other cases the
converse was found. The general findings of the research are as
follows:
750
Advantages of Using QuickBird:
* major shapes of all buildings can be identified;
* demolitions are casy to detect;
* small industrial buildings are easy to identify;
* QuickBird imagery is useful in areas where access is
restricted;
® the imagery provides wide-area coverage and fre-
quent repeatability ;
e the imagery provides a time-stamped snap-shot of the
currency on the ground.
Disadvantages of Using QuickBird:
* QuickBird imagery can’t be used to identify whether
change is permanent or temporary;
the age of changed features cannot be identified;
* the Category of the change ( A or B), is difficult to
determine, especially for urban areas;
® drive restriction features are difficult to identify (e.g.
speed bumps and posts);
* — high-rise buildings are difficult to determine in the
images, and any overthrow makes it difficult to iden-
tify if any change has occurred;
* complex shapes and multilevel structures are very dif-
ficult to see clearly;
* image resolution is often not high enough to differen-
tiate between building extensions and separate new
buildings;
e all small linear features are difficult to see (e.g.
fences, paths, railway lines or field boundaries).
The map currency audit must also identify any cultural or
thematic changes, such as changes of use or changes in names
of features. These of course cannot be obtained from imagery.
As in the change detection example, the results show that
QuickBird imagery can be used in the map currency audit
process, but it cannot be used alone. Additional information is
required which can only be obtained by a field visit to the sites
under investigation.
4. CONCLUSIONS
The results described in this paper have indicated that
QuickBird Imagery can play a role in all the processes
investigated.
Imagery of this resolution can be used to update mid-scale
maps (1:6 000 to 1:10 000 scale) as long as small linear features
are excluded from the mapping specification. This imagery can
also be used in the detection of change, and in the quality
checking of existing map data. However, in each case there are
disadvantages, which indicate that QuickBird imagery should
be used in a supplementary way, rather than as the main source
of data. For example, QuickBird could be used to obtain
frequent snapshots of rapidly changing urban areas, enabling
change to be detected more readily than is possible by other
methods.
S. REFERENCES
European Space Imaging, 2003, “Flanders Chooses IKONOS”,
press release,
htip/www.euspaceimaging.com/content/Downloads/news/15-
july-2003.pdf (Accessed 28th April 2004)
Internatic
Holland,
Topograr
OFEPE
Holland,
Imagery i
ISPRS/E;
Mapping
http://www
hannover
(accessed
Kumar, ^
Imagery 1
Mandevil
March/Aj
http//ww
28th Apri
Petrie, G.
Monitorir
January/?
Ridley, H
Dowman,
commerci
Ordnance
Sensing, €
