International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004 
in the very recent past has DBMS technology been available 
to support such extended continuous spatial coverages. 
The need for consistency across joined-up data arises also 
when topographic framework data is brought together with 
similar data from other adjacent jurisdictions or with other 
layers of business data. 
2.3 Conflation and use of best Available Sources. 
Imagery is becoming increasingly important as the primary 
source for topographic framework data. According to 
Heipke, it will account for around 50% of such data in most 
NMA's over the next few years (Heipke, 2004). Other 
sources are also important including field survey and other 
data. An enterprise-wide database-centric approach is needed 
to support exploitation of this rich set of sources. Integration 
of photogrammetry with the database is part of this. 
A key element for improved efficiency in the future is 
conflation — the combination of two datasets to produce a 
merged dataset with the best elements of the inputs. 
Automation of conflation depends on the articulation of the 
rules for determining the required best elements, and their 
implementation in a rules-based processing environment. 
Such rules will likely be imprecise and data will never quite 
fit, so fail-safe recourse to human interpretation is needed. In 
future this kind of approach will increase the degree of 
automation in data update and enhancement using imagery as 
well as other sources. 
3. LEVELS OF INTEGRATION AND OF 
INTEROPERABILITY 
3.1 File-based Data Exchange 
Most implementations combining photogrammetry with GIS 
databases in use at present use file based-data exchange with 
the database. Data is extracted, updated and returned, either 
on a whole replacement basis or in some cases as files of 
deleted and created features. The data model is essentially a 
simple feature model. A variety of proprietary CAD or GIS 
formats are used (DGN, DXF, Shape) and often there are 
problems arising from loss of information (eg multiple 
attributes). Some organisations have developed lossless 
exchange formats. Validation of the modified data takes 
place at the end of the session(s), on return to the database. 
Failures of validation result in repeated revision cycles, often 
over long timescales if the validation processes are 
prolonged. Nevertheless this architecture is simple to realise 
and widely adopted. It represents the initial level of 
integration and can of course support remote operation. 
760 
   
    
  
Revised Extract 
or Changes 
Extract 
Validation 
    
  
Spatial DBMS B 
   
Fig. 1 File-based Data Exchange. Several cycles over 
extended timescales. 
3.2 Direct Link to GIS Database 
The next level of integration is achieved by a direct link 
between the photogrammetric system and the GIS database, 
by programmatic connection using the APPs of the two 
systems. This avoids problems of information loss. Instead of 
a single commit to the database, there is a series of commits 
on a per completed operation basis. There is some gain in 
efficiency due to the closer coupling and a reduction in the 
'floppynet effect. Most mainstream photogrammetry 
systems can now operate in this mode with a GIS database 
such as ESRI's ArcSDE or via a GIS layer to Oracle Spatial. 
  
Interna 
  
Fig. 2 
3.3 Di 
Gains ii 
and/or « 
validati 
expande 
active \ 
efficient 
into us 
processi 
example 
provide: 
(Edwarc 
At the 
manage 
technolc 
technolc 
coupling 
validatic 
cycle op