The interpretations were actually to a great extent 
based on comparing the old aerial photos (since the old 
map does not contain clearings) to the new ones and on 
the knowledge of the average growth in the area. It was 
estimated that the length of time from cutting to 1.5 m 
was the same as between the old and new images. 
For a good result in future revision interpretations, it is 
of great importance to know the approximate age of the 
clearings, since the age is somewhat correlated to the 
tree height. 
A further analysis of misinterpretations shows that 
some very small clearings were left out in some of the 
interpretations (although they were large enough to be 
represented in the map). Another fact was that a 
number of stands that were clearings in the old map, 
still were clearings in 1989/90, because the renewal in 
the test area was not as quick as it was assumed to be. 
4.7 Clearings - geometric accuracy 
Figure 1 shows one clearing interpreted in different 
image materials. No significant difference in the geo- 
metric accuracy of the delimitations, of importance to 
this application, is found between the materials. 
4.8 Built-up areas 
Experience shows that delimitations of new built-up 
areas in SPOT images can be interpreted while the 
classification into different categories of urban areas is 
not possible. For aerial photos the situation is much the 
same. 
4.9 Time 
The amount of time spent on the different interpreta- 
tions is approximately the same. 
4.10 Costs and accessibility 
Costs and accessibility for SPOT data vs aerial photos 
were evaluated for the applications in this study. 
Costs for air photos from 9 200 m are higher than for 
SPOT data. Air photos from 13 200 m are however less 
expensive than SPOT data. As an aim at the National 
Land Survey to develop more efficient methods for map 
production, digital interpretation methods are of great 
interest. Costs for satellite data benefit from the use of 
digital methods. 
Accessibility factors are more difficult to compare. Both 
SPOT data and aerial photos should have the potential 
of supplying NLS with the needed data for the map 
revision application studied here. For SPOT data, 
programming demands from other customers could 
cause a conflict. Also, extreme weather situations, 
especially in the very northern parts of the country, can 
influence the outcome from both the SPOT satellite and 
from aerial photography. 
5. CURRENT WORK 
Current work involves evaluation of digital revision 
methods with SPOT data at the National Land Survey. 
561 
The geographic information system ARC/INFO is one of 
the digital systems used for cartographic purposes at 
the National Land Survey. The ARC/INFO workstation 
version 6.0 software allows raster images, such as 
satellite imagery, as a background to geographic data 
on the display. The satellite data can be contrast 
stretched to enhance certain features of current in- 
terest. The raster data is registered to the map data so 
each point corresponds to the same point on the ground. 
It is possible to zoom and move around in the images. 
À future revision process using SPOT imagery could 
typically include the following steps: The existing 
digital data (clearings, roads, etc.) is displayed on the 
screen. Newly acquired satellite data, delivered on 1/4" 
cassettes, is also displayed. New features in the image 
is digitized using ARCEDIT and eventually used to 
update the objects in the database. 
The described process is an available technique today at 
NLS, made possible through todays development in 
hardware and software for geographic information 
systems, and also because the topographic map data at 
NLS is digitally stored. 
We believe that a future step in development of digital 
revision methods is semi-automatic techniques taking 
advantage of both the inherent automatic interpretation 
possibilities in digital satellite data and the inimitable 
capability of human decision involved in image in- 
terpretation. 
6. CONCLUSIONS 
This study shows that, despite differences in geometric 
resolution in the image materials, no differences of 
significance for topographic map revision are found in 
the interpretations of roads and clearings in air photos 
(9 200 meters), SPOT XS or SPOT PAN, regarding the 
different accuracy measures: completeness, shape and 
geometric accuracy. However, roads on the actual clea- 
rings have been found to cause some problems in XS. 
In SPOT images the delimitations of new urban areas 
can be interpreted while the classification into different 
categories of urban areas is not possible. For aerial 
photos the situation is much the same. 
The result also indicate that training is needed to fully 
take advantage of the multispectral images of SPOT. 
Costs for air photos from 9 200 m are higher than for 
SPOT data while air photos from 13 200 m are less 
expensive than SPOT data. However, for digital revision 
methods, which is likely to be used in the future, 
satellite data has the advantage of being in digital form. 
Accessibility factors are more difficult to compare. Both 
SPOT data and aerial photos should have the potential 
of supplying NLS with the needed data for the map 
revision application studied here. 
The conclusion of the study is that SPOT data is a 
useful data source, for topographic map revision at the 
scale of 1:50 000 in Sweden, regarding roads, clearings 
and built-up areas.