S system 
s Ltd. of 
a. CARIS 
S system, 
abilities to 
nyi, 1991; 
hic digital 
S software 
ta capture, 
tion and 
creation, 
a analysis, 
nated map 
dressed in 
e analysis. 
'ommands 
the colour 
age, select 
> selected 
nages and 
'etation is 
hancement 
n or user 
ar stretch, 
lirectional 
egistration 
ng, ortho- 
sification 
ot project 
on 4.24 
| 10. 
'S MAPS 
ct were to 
] data set, 
a set, and 
using the 
1:250 000 
ing to the 
sen as the 
area also 
coverage. 
revised in 
ard copy 
imagery 
sheet was 
‘a current 
ılts of the 
with the 
materials 
were the 
ital data, 
  
  
reference points, digital Landsat TM imagery, 
aerial photography and the NTS maps at both 
1:50 000 and 1:250 000 scales. 
3.1 Derivation through the Generalization of 
the 1:50 000 Digital Data 
For this project the vector data files were loaded 
in SIF format and converted into the CARIS 
data files required for processing and display. 
Based on the review of the procedures currently 
being used for the manual derivation of a 
1:250 000 NTS map from 1:50 000 maps, the 
methodology of deriving one layer at a time was 
used to digitally derive the pilot area. The 
following sections describe the steps of the 
derivation of each layer. 
3.1.1 Review of the 1:250 000 Specifications 
To determine the required content and minimum 
size for the derived data set, the 1:250 000 
Polychrome Mapping Specifications were 
reviewed. Content refers to the features that are 
shown on the 1:250 000 map and minimum size 
refers to the smallest area or line to be shown. 
To complement the specifications which are 
currently being revised, the project team relied 
quite strongly on the directives and minimum 
size guide being used in the Map Revision 
Section. 
3.1.2 Data Generalization 
The generalization process used for the 
derivation of the 1:250 000 data set from 
1:50 000 data consisted of the following eight 
techniques (Mackaness, 1991): 
a) Feature correlation/selection: 
Having identified the required content, the two 
data sets were first correlated in order to match 
corresponding features. A direct correlation of 
feature codes was not possible because the 
feature codes for the 1:250 000 data are not all 
the same as those used for the 1:50 000 data. 
Next, the content required for the 1:250 000 
data set was extracted from the 1:50 000 data set 
Finally, the feature codes of the selected data 
were changed where necessary. Caution must 
be taken during this stage not to delete any 
features that are required for data continuity, 
such as dugouts that are located on a river. The 
correlated data was separated into the layers; 
hydrography, transportation, vegetation and 
wetland, culture, built-up area, contours, and 
gravel pits, and layer numbers assigned. 
b) Data filtering: 
The Douglas-Peucker filter algorithm was 
applied to the transportation layer at 50m and to 
the hydrography layer at 25m intervals. These 
filtering intervals were tested and chosen to 
ensure that the data retained its shape while 
reducing the amount of data points, which in 
turn accelerated the data handling and 
processing. 
c) Omission of features not meeting minimum 
size requirement: 
Using the minimum size guide, features that 
were too small, too narrow, or too close 
together were identified and interactively 
removed from the data set. The exception being 
those features that could be grouped together or 
that were required for data continuity such as 
transformer stations on power lines. 
d) Grouping of features: 
Polygons not meeting the minimum area 
requirement were interactively grouped together 
or buffers were generated surrounding them. 
e) Collapse of features: 
The collapsing technique was first used when a 
feature did not meet the minimum size 
requirement to be shown as one type (i.e., an 
area) but could be represented as another type 
(i.e., a point). In the cases of a double line 
river that did not meet the minimum size, a new 
line was digitized representing the river center 
line, and a campground area that was not large 
enough to satisfy the minimum requirement was 
deleted and replaced with a point type. A 
second situation where collapsing was required 
was where a feature, such as an interchange 
was represented as one type at the 1:50 000 
scale, but as another type at the 1:250 000 scale. 
f) Combination (reclassification) of features: 
Reclassifying features for generalization means 
changing their feature code in order to connect 
the feature with an adjacent feature. Segments 
of lines not meeting minimum length 
requirements such as ditches at the end of a 
stream were reclassified to stream. 
g) Simplification (line smoothing filter) of data: 
Line smoothing was applied to features that 
were jagged to make them cartographically 
presentable. It was necessary to smooth the 
streams from the 1:50 000 data for presentation 
at the 1:250 000 scale. 
289