ting 
ting 
the 
gely 
the 
ve a 
rf 
king 
the 
Ss, a 
and 
here 
able 
ntil 
into 
ital 
ence 
uced 
ence 
ated 
etry 
S iin 
rof 
ller 
rate 
ctly 
the 
ntly 
the 
ion. 
tems 
, tof 
deal 
tion 
ired 
age, 
sary 
and 
the 
pear 
ar of 
ical 
croi 
updating environment to establish if 
this really is the case. A’ lesser 
Specification could well produce 
acceptable results in a monoplotting 
System, depending on the nature of 
the area under revision. 
Digital Elevation Model (DEM) 
Here again the accuracy of the 
resultant digital  orthophoto ig 
influenced significantly by the 
quality. of «the. DEM used in . its 
generation. Indications are that 
contour information, for example, 
would need to be in the order of one 
to five metre vertical interval to 
produce an orthoimage of suitable 
accuracy for large scales revision 
work. However while this may be 
necessary in areas of large height 
variation it may be possible to get 
by in a lot of less undulating areas 
with a more relaxed vertical interval 
such as 10 metres. 
Alternatively a suitable DEM can be 
produced automatically using 
previously generated height control. 
However there are indications that 
editing of some 20$ to 30$ of points 
could be necessary and this requires 
a stereoworkstation facility. 
Obviously there are aspects here 
which require greater examination in 
terms of the large scale updating 
task if the less expensive monoscopic 
route is pursued. 
Whatever method is used quality of 
the resultant digital orthophoto is 
fundamental to its successful use 
with superimposition of the vector 
database. This is doubly so as the 
orthophoto is a marketable product in 
its own right. 
Updating from the Digital Orthophoto 
Given... a. .digital.. orthophoto - of 
suitable quality then superimposition 
and temporary warping of the vector 
database for consequent updating is 
relatively straightforward in 
principle. However in most cases a 
building for example will be seen in 
Some sort of perspective view with 
37 
perhaps two sides visible at ground 
level, zooming in with accompanying 
blurring of the raster image, coupled 
with shadow, can make for quite 
difficult accurate interpretation of 
features for update. Further 
assessment of the inherent problems 
here is desirable. 
Databasing 
While | updating of topographical 
features can be considered feasible 
the degree to which such work can 
accommodate the attribute aspects of 
the updates is where the possibility 
of significant benefits exists. Given 
complex large scale, topologically 
structured, multi-feature coded data 
requiring update there are a lot of 
issues remaining to be addressed if 
the monoplotting revision option is 
to be really cost effective in the 
widest sense. 
Other Considerations 
Particularly in stages (ii) and (iii) 
above the major concerns are to 
produce a digital orthophoto which 
will enable revisions to be extracted 
to the required accuracy. Further 
assessment of the potential of 
monoplotting for large scales update 
must also examine an even more 
radical scenario. This involves use 
of a straight forward digital 
representation of the aerial 
photograph to provide a sufficient 
revision answer over a localised area 
when the vector database is overlaid 
and warped. to, £36. If 1% works it 
will certainly be cost effective but 
it must be proven to produce the 
required accuracy. 
CONCLUSION 
Considerable development 1s still 
needed to produce really efficient 
updating procedures for complex large 
scale digital topographic databases. 
Significant progress is evident and 
fuller handling of the complexity 
aspect at data capture stage appears 
to offer the best future line of 
research. This applies particularly 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B6. Vienna 1996