1 
| 
Hi 
Ei 
Li 
H 
  
La ir ee ES EEE 
Several formats have been developed for DIM sur- 
faces. Some of the main formats used in computer 
software are: 
1) Grid Format (GRD) 
A set of x, y, z digital coordinates that 
represent the terrain as a regular, rectangular 
pattern in the x-y plane. 
Triangle Format 
A set of x, y, z digital coordinates that 
represent the terrain using a pattern of equi- 
lateral triangles in the x-y plane. 
3) Triangle Irregular Network(TIN) 
À triangle network which represents the terrain 
using triangles of variable sizes and shapes 
in the x-y plane. 
4) Topological Triangle Network (TTN) 
A form of TIN network which also represents the 
terrain with spatial BREAK lines, OBSTACLE 
lines, SPOT elevations, and REGULAR points to 
show spatial location and topographic relief. 
2 
— 
A TIN file is a data file used to store the Topo- 
logical Triangle Network data. This file contains 
all of the input terrain data and the resultant 
triangle network converted from the topological 
geographic elements. It is generated according to 
a designed data structure which uses attributes to 
store related information about each triangle, 
neighboring triangles, vertical scale, elevation 
datum, multiplier and other information. 
REGULAR points form triangles under the Delauny 
Condition which states that when all triangles in 
the surface have been formed no other REGULAR point 
should be contained within a circle prescribed by 
the three vertices of any triangle. This condition 
is overridden in the vicinity of BREAK lines, 
CONTOUR lines, OBSCURE areas, SPOT, and EDGE fea- 
tures in order that no triangle side crosses these 
features. 
For a large project, such as a highway or landfill 
design, the DTM file is usually created with 
three-dimensional geometric elements digitized with 
a stereoplotter and converted into the Topological 
Triangle Network (TTN) Format. The DTM file in TIN 
format should be, theoretically, more accurate and 
more practical than other formats. The TTN format 
can represent a broad area with a limited number of 
points defined by variable densities according to 
the relief of the terrain and the required accuracy 
of the application. Another advantage is that only 
the TTN format can model accurately the critical 
high and low positions, such as road shoulders and 
ditches, without any additional efforts. 
The Intergraph InRoads software, a TTN based DTM 
modeling and design package, was used to implement 
experiments which try to find the most efficient 
method to create an DTM file that meets engineering 
accuracy requirements. A study area of approximate- 
ly 2100x800 feet was selected for these trials. 
The photogrammetric data collection was performed 
using a KERN PG2 analog stereoplotter interfaced 
with an Intergraph workstation. 
All the geographic elements of the DTM were digi- 
tized by the same operator with the same stereo- 
plotter in order to reduce the interference of 
human and system errors. These DTM files were then 
used for creating contours, profiles, cross-sec- 
tions, and for estimating cut and fill volumes for 
a highway design. 
844 
The standard procedures implemented in this experi- 
ment were as follows: 
1) Set up the stereo model and design file; 
2) Digitize EDGE line (boundary) on a specific 
level with corresponding symbology; 
3) Digitize BREAK lines, REGULAR points, and SPOT 
elevations on different level and different 
symbologies; 
4) Digitize random contours for verification 
purposes; 
5) Convert 3D elements to TTN file; 
6) Generate and display contours from the stored 
TTN file; 
7) Compare the contours generated from the TIN 
file with the verification contours. 
According to our experience, the REGULAR points 
take most of the digitizing time. Because the 
operation time is proportional to the total amount 
of digitized points, the number of REGULAR points 
digitized has the greatest impact on the cost. For 
easy comparison, we digitized four files using the 
same BREAK lines, SPOT elevations, and EDGE lines, 
but different densities of REGULAR points. For the 
convenience of digitizing, we used the vertices of 
3D line strings to represent the REGULAR points. 
(see Fig.1) 
The final contours were generated at one foot 
intervals based on the DTM file with the following 
specifications: 
1) 25 FT REGULAR Spacing - See Fig.2 
2) 40 FT REGULAR Spacing - See Fig.3 
3) 50 FT REGULAR Spacing - See Fig.4 
4) 75 FT REGULAR Spacing - See Fig.5 
3. ANALYSIS and COMPARISONS 
The basic assumptions used in the studies and 
analysis were as follows: 
1) This is a relative comparison between the 
contours generated from different densities of 
geographic points and manual stereocompila- 
tion. 
2) We are presuming the digitization to be accu- 
rate. 
3) The elevation of an individual point is more 
accurate than the elevation of the digitized 
contours. 
4) Assume the final map scale is 1:600, and it 
will be used in a highway project. 
A field verified survey is needed to evaluate the 
error in the contours from manual stereocompila- 
tion. Without a field verified survey, we assume 
the manual compilation meets National Map Accuracy 
Standards. 
Table.1 is a list of elevation differences between 
the contours generated from the DTM data with 
different specifications and the contours compiled 
manually. The differences are more uniform as the 
spacing becomes smaller. 
Table.2 lists the number of points for each geo- 
graphic element in different DTM files. The collec- 
tion time and standard deviation for each file are 
also shown. These data are used to create Fig.6 for 
determining the optimal number of points to col- 
ect. 
VM BB COUT = OS Or 
cC» Se oc — = 
—. TIO + Wn