6. INTEGRATION OF GIS WITH 
RASTER PROCESSING 
A number of other functions integrate vector 
data with raster data in general. These are 
important as they increase the general 
capabilities of GIS systems for handling raster 
datasets. The most common processing 
functions are: 
6.1 Raster editing. 
This is becoming increasingly important within 
a GIS, as raster base maps are used as a cheap 
and quick alternative to vector digitising the 
maps. Raster editing functions required 
include: 
e Outlining areas to be edited by a vector 
polygon 
Creating new vector objects using 
sophisticated drawing tools (such as text 
font, positions, sizes, orientation), and then 
"burning through’ into the raster dataset. 
These techniques are increasingly being 
considered by map publishing agencies either 
as an alternative to a fully vector database, or 
as the basis of a hybrid, evolutionary approach, 
spreading the costs of data capture in a 
manageable way. 
6.2 Raster modelling. 
This is the term used for the combination of 
two or more raster datasets using mathematical 
(+-*/)) and boolean (AND, NOT, OR XOR) 
operators. More sophisticated packages include 
functions such as cost surfaces and zone 
operations (eg. find the perimeter of a zone of 
values), and have constructs such as DO and 
WHILE loops. These have the potential of 
integrating vector data within the same type of 
analysis constructs, such as ’IF (within 
polygon) AND (attribute = X) THEN... 
Although the output data structure will, in this 
case, still be a raster, vector data can be 
handled in the same way as a raster. 
388 
The ability to model using multiple rasters, 
where raster data can be from a variety of 
remotely sensed image bands, DTM and 
scanned map sources facilitates exploration of 
complex modelling scenarios such as 
estimation of cross country vehicle movement 
potential in fragile environments where limited 
conventional mapping may be available. 
7. A STEP FORWARD 
Real advances in the integration of remote 
sensing and GIS can only come with the 
development of integrated software 
environments where the display, interrogation, 
processing and analysis of raster and vector 
datasets is supported by a single system 
architecture which gives freedom to model and 
rapidly build applications according to the task 
in hand rather than being constrained by the 
functionality of a fixed function GIS or image 
processing system. The underlying database for 
such a system should be capable of integrated 
storage and indexing of any GIS or remotely 
sensed data, with a unified data access 
mechanism. Only with ready access to such 
heterogeneous data, modelling opportunities 
can increase in complexity. 
Using the latest object oriented GIS technology 
Laser-Scan have developed such an integrated 
GIS and remote sensing system for the British 
National Space Centre (BNSC) under a 
contract placed and managed by the Defence 
Research Agency (DRA) Farnborough. Called 
IGIS, the system is built around a versioned 
object database which provides an intuitive 
way of holding spatial and non-spatial data. 
Both raster and vector continuous maps are 
supported. The object model provides a natural 
correspondence between real-world objects and 
the data that models them. As well as support 
for a variety of datatypes, IGIS also supplies 
references which enable objects to refer to each 
other. Collections of objects, with raster and 
vector components, can be combined into 
complex structures which mimic those found in 
the real world. Navigation of these structures 
provides a natural 'distributed' index enabling 
rapid access between related objects. 
In addi 
(‘metho 
mechan 
gives er 
of com] 
comple» 
8. E 
A] 
The int 
section 
Departn 
funded 
running 
: Council 
test the 
give La 
the syst 
that the 
needs of 
Each p 
require 
sensed 
The di 
flexibili 
e BNS 
Devi 
inves 
SAR 
in 
parar 
of | 
locat 
irreg 
homi 
defir 
imag 
publi 
editi 
spati 
boun 
(for: 
girth 
vecti 
a ca 
spati 
remc 
requ