DLG 
imes 
urce 
ion. 
iate 
- ion 
tric 
tric 
, 000 
ass" 
1 is 
late 
lata 
hen 
.ant 
; Of 
3 Or 
Iser 
ICY 
lata 
on, 
the 
the 
oto 
ved 
map 
OW. 
ple 
fic 
one 
in 
een 
the 
and 
it. 
old 
no 
nth 
ric 
out 
ole 
che 
Lth 
ite 
Feature Completeness Accuracy 
Feature completeness is used in this context 
to define the amount of information contained 
in the map graphic data set. For image 
mosaics it is directly related to image 
resolution and for line graphic data it is 
related to the original compilation scale. 
Regional and local users will have different 
requirements for information content. This 
is a topic that creates a situation of weights 
and balances, based on the data handling 
capabilities of the geographic information 
system. Typically a vector data set over a 
specified area takes much less computer 
storage space than a raster image data set for 
the same given area. A vector digital map 
graphic data set is comprised of only selected 
features of interest, consequently only those 
features are available for the users 
applications. For the DLG databases that 
would include, transportation, hydrography, 
selected political/administrative boundaries, 
and contour lines. However if you have an 
raster image map graphic data set, you have a 
comprehensive view showing everything visible 
to the sensor for the given area. With SPOT 
imagery the information visible to the sensor 
equates to the 10 meter pixel resolution, so 
that anything smaller than 10 meters may not 
be detected by the sensor, unless it contains 
a significantly high tonal contrast with its 
surrounding area. For this reason SPOT 
imagery does not meet all of the needs of a 
local user group. However as discussed in the 
applications section of this paper, there are 
definitely uses for SPOT imagery for the local 
users in terms of query and location specific 
applications. 
Feature identification is another important 
part of the feature completeness accuracy. 
Significant roads and buildings, along with 
all natural resource features such as lakes 
and forest areas, are easily recognizable on 
the SPOT images. Problem areas exist where 
vegetation covers roads or buildings. But 
this is a problem more significant to local 
users. There are also many situations where 
determining the significant shape of the 
building is not as important as knowing that 
it exists and its relative location. A large 
house for instance 60’ by 75’, will be 
reflected in 6 pixels making it easily 
identifiable in an image. A small house 30’ 
by 30’ will be reflected in 1 or 2 pixels, not 
significant in terms of identification, but 
allowing the user to see that there is indeed 
a building at that location. FIGURE 1 shows 
that many features on the image are quite 
easily detected, including many of the 
residential dwellings. As noted above it may 
not be possible to assign attributes to all 
those features. 
The advantage of the DLG data is that each map 
graphic data item has already been selected, 
classified, and symbolized. With SPOT mosaics 
you do not have symbolized information. 
However, you do have a much more comprehensive 
view of the existing ground information which 
allows you to create your own feature 
identification and classification categories 
with information that is not present on the 
DLG map graphic data sets. 
Accuracy Summary 
To summarize the accuracy issues, each of 
these accuracy types relate to regional type 
databases or data sets that currently exist. 
TABLE 2 compares 1:100,000 USGS DLG data, 
1:24,000 DLG data as it would be digitized 
from a USGS 7.5 minute quad sheet, and 
mosaicked SPOT imagery. The rectified SPOT 
image mosaic carries a higher metric accuracy 
than the 1:100,000 data, and approaches the 
metric accuracy of the 1:24,000 DLG potential. 
Also significant is that when you look at the 
effort involved in combining graphic data, 
such as multiple DLG files or multiple SPOT 
scenes, the differences in metric accuracy 
begin to look much more significant. 
For temporal accuracies the SPOT imagery 
provides a much more useful solution. Not 
only can you get updated information in a 
relatively quick time frame, but you can 
develop an initial data set that is more up to 
date than the USGS DLG data. Unfortunately 
the 1:100,000 DLG data that exists today, in 
many cases is significantly out of date. 
For feature completeness, the SPOT imagery is 
a comprehensive view of the existing ground 
information. The DLG data provides a data set 
in which selected features such as roads, 
waterways, and administration boundaries, 
along with topographic contours and names are 
identified for the user. For the user 
interested in only one or a combination of 
those features included in the DLG, it is a 
much easier data set to use. All of the 
information has been classified and 
symbolized. For users needing a comprehensive 
view of the area, the mosaicked SPOT imagery 
becomes an important map graphic database. 
TABLE 2 
DATABASE COMPARISONS 
Positional Coverage 
  
Maps or Images / Accuracy 
SOURCE Accuracy Per map/image (sg. miles) 
(feet) (sq. miles) 2500 125,000 250,000 
SPOT 50 1400 2/50' 90/75' 180/100’ 
DLG 100 165 800 4/ns 156/ns 312/ns 
DLG 24 40 50 50/ns 2500/ns 5000/ns 
  
*ns = not specified