2.2 DEM Corrected Tracing 
This scheme requires a digital elevation model (DEM) in 
either a grid or triangulated irregular network (TIN) form. 
Based on this data, the relief displacement of points along 
the path of digitization is determined and their planimetric 
position corrected. 
Interpolation in the DEM is performed according to an 
iterative algorithm developed by Masry and McLaren (1979) 
for computing the planimetric coordinates of features 
digitized in a single photograph. First the exterior 
orientation of the photography is acquired. Next, the 
collinearity equation is used to project the image points onto 
a plane set at the average terrain elevation to obtain object 
space coordinates X', Y'. The corresponding terrain 
elevation Z' is then found in the DEM which in turn defines 
a new projection plane. The procedure is repeated until the 
change in Z' is within a preset tolerance. The X, Y, Z 
values obtained at the last iteration are adopted as the object 
space coordinates of that particular image point. 
2.3 Digital Orthoimage Tracing 
The acquisition of planimetric object space coordinates from 
monoscopic images becomes a simple task if both the tilt and 
the relief displacements have been removed from the image 
prior to digitization; in other words, if the feature tracing is 
performed in an orthoimage. 
Differential rectification of raster imagery is a straight- 
forward process. First the exterior orientation parameters 
are acquired. Next an empty grid is created at the proper 
orientation and with a spacing that corresponds to the desired 
pixel size of the orthoimage. The elevation of each grid cell 
is then interpolated in the DEM, and is combined with the 
exterior orientation parameters in a transformation of the 
empty grid onto the uncorrected image. Finally, the gray- 
scale value of each grid cell is determined by one of the 
resampling algorithms such as the nearest neighbour, 
bilinear or bicubic interpolation. As a byproduct of this 
scheme, the orthoimages can also serve as a companion to 
digital line maps. Images are a more complete representation 
of the real world than line maps. 
3. MAP REVISION TEST 
3.1 Test Material and Preparation 
All three revision schemes were tested on a 1: 50 000 scale 
digital map sheet of the Canadian National Topographic 
System (NTS) series and on a 1: 10 000 scale digitial map 
sheet of the New Brunswick Geographical Information 
Corporation (NBGIC). To facilitate the evaluation of the 
results, various features were deleted in a copy of each map 
file to simulate an out-of-date map. These features were then 
re-established from the same black and white aerial 
photographs used for the original map compilation. The 
photo scale was 1: 40 000 and 1: 35 000 respectively. 
The digital images were obtained by scanning the paper 
prints in a Hewlett-Packard ScanJet Plus document scanner 
at 300 dots per inch (118 dots per cm) which resulted in a 85 
um pixel size at image scale. The corresponding pixel size 
on the ground was 3.4 m and 3.0 m in the two photographs 
respectively. The radiometric values were recorded in 256 
gray levels. 
In preparation for the piecewise rectified digitization and the 
DEM corrected tracing, the image was registered. to the map 
by perspective transformation. The orthoimage was 
produced by a rigorous pixel-by-pixel differential 
rectification, based on the collinearity equation. The cubic 
534 
convolusion resampling was employed in all cases. Road 
intersections and other well-defined features were selected in 
the screen display of the digital maps as control for the 
transformation. 
Both maps cover the City of Fredericton and vicinity. The 
downtown area which spreads along the shore of the St. 
John River is essentially flat ground at an elevation near sea 
level. There from the terrain has a steady incline and reaches 
an elevation of 130 m at the city limit. 
3.2 Data Collection and Evaluation 
Change detection was performed visually on the screen in a 
merged map and image display. The new features were then 
traced in the image by freehand cursor control and digitized. 
The features mapped included: highways, major 
thoroughfares, residential streets, river shoreline, power 
transmission line, racetrack, highway bridge, buildings and 
edge of forest. All features were digitized in point mode. 
Roads were digitized along the centre line. 
The new features digitized in each of the three revision 
schemes were concentrated into the original map. The 
position of the features in the original map were used as the 
reference to ascertain the accuracy of the newly digitized 
values. Point features were tested by forming the coordinate 
differences. Line features were subdivided into sections at 
well identifiable breakpoints. Thereafter, X, Y coordinates 
were generated at equal intervals along the original and 
newly digitized path of the features. The deviations at 
corresponding point pairs were then computed. 
Final assessment of the three map revision schemes was 
based on the degree of compliance with the map accuracy 
standards. The 1:50 000 NTS map was produced to meet 
the "A" rating in the classification as to accuracy of 
planimetry ("Circular Map Accuracy Standard") in the North 
Atlantic Treaty Organization (NATO) Standard System for 
the Evaluation of Land Maps, which states: 
Planimetric position of 90% of well-defined features 
measured from the map (except those unavoidably 
displaced by exaggerated symbolization) will fall 
within... [25 m]... relative to their true planimetric 
position as referred to the geographic graticule or 
grid of the map. 
This accuracy criteria expressed in terms of measurements 
made on the map is 0.5 mm. The equivalent root mean 
square error (RMSE) in position is 16.5 m or 0.33 mm. 
The 1:10 000 scale map was produced by the Maritime Land 
Registration and Information Service to satisfy the accuracy 
specification for the Urban and Resource Digital Map Base 
prescribed in the Land and Water Information Standards 
Manual, which states [NBGIC, 1991]: 
Ninety percent (9096) of all "well defined features" 
must fall within the positional accuracy... [2.5 m]. 
Well-defined features are those whose positional 
accuracy is not adversely affected by vegetative 
cover. Accuracy of the digital data (point, line, area) 
can be defined as the difference between the position 
of the associated data in the digital file and the real 
position of the represented feature on the earth. 
The above requirement expressed in terms of RMSE in 
position is 1.6 m. It should be noted, that the accuracy 
assessment was based on the comparison of two digital 
products and not on measurements made on the ground. 
Therefore, the tolerance can be increased by a factor of V2 to 
account for the uncertainties in both data sets, which is equal 
to 3.5 m and 2.3 m for the 90% error and the RMSE 
respectively. The required 9096 error and RMSE for a 
graphical plot produced from the digital map base are 5.0 m 
and 3.3 m respectively, or 0.5 mm at the publication scale.