239 
This system is designed to take in multiple 
imagery sources, including digitized aerial 
photographs, to extract from this imagery 
geographic information with a high level of 
positional accuracy, and to output this data 
to a variety of GIS systems via standard 
exchange formats (see Figure 3). The 
information extraction process, which 
includes planimetric data as well as 
elevation, is automated wherever practically 
possible, and the interactive operator 
quality control process is aided by a high 
resolution stereo display station. The 
window based operator interface features 
fully integrated image and graphics 
manipulation. The system is being designed 
for extensibility by using object oriented 
desiqn principles, and features a local 3-D 
data base (extensible to 4-D) providing for 
coordinated rapid storage and retrieval of 
related image, graphics and attribute data. 
The development of Geomate is being carried 
out by MacDonald Dettwiler and is sponsored 
by the Ministry of Energy, Mines and 
Resources in Canada, who will be the first 
user. Tools like Geomate will encourage 
increased use of remotely sensed data in 
populating and updating geographic data bases 
for a variety of applications. As these 
systems become available over the next few 
years, another factor limiting the growth of 
remote sensing applications will diminish. 
Data Base Structures 
Similar to processing tools, the data storage 
structures available at present are generally 
geared towards a single problem domain. 
While reasonably efficient structures are 
available for 2-D image data, 2-D vector data 
and attribute data individually, and to some 
extent for vector and attribute data 
combined, the application of remotely sensed 
imagery will require better mechanisms to 
maintain historical and logical relationships 
between the various data types. In addition, 
the third dimension will have to become more 
fully integrated for topographic applications 
to work well. Similarly, means to deal with 
the time dimension must be devised to allow 
rapid dynamic modelling of the data, an 
essential capability for effective resource 
and environmental management applications. 
Finally, the proliferation of computer power 
throughout organizations today drives the 
need for a combination of central data 
management and the ability for workstations 
to safely operate on authorized portions of 
the data. 
All of these problems have been known for 
some time, and a certain level of agreement 
now exists that the object oriented design 
technologies now available may enhance our 
ability to address these problems. So far, 
however, the software industry has not yet 
come to grips with this problem domain. 
Recognizing this situation, Canadian 
Government and industry made it an integral 
part of the Geomate development carried out 
by MacDonald Dettwiler to make progress in 
this area. 
Further advances in this area are expected to 
come out of the data base research labs over 
the years to come. This then will reduce yet 
another factor which has contributed to the 
under-utilization of remotely sensed data, by 
providing data structures suitable for the 
dynamic modelling process which should 
precede resource and environmental policy 
formulation. 
Affordability of Computer Technology 
Due to its very nature, the efficient 
handling of remotely sensed data requires 
considerable computer processing and storage 
resources. Typical systems in years past 
Figure 3 Hie Geomate Concept featured powerful mini-computers with array 
processors, expensive image display 
processors, and a bank of disk drives. 
Multiple workstations often required 
repetitive inclusion of these expensive 
devices. The price for acquisition and 
operation of these systems could only be 
afforded by a relatively small group of 
users.