International Archives of the Photogrammetry, Remote Sensing
points; thus when ever the surface modeling factor is not
reached and there are no obstacles, data are to be collected in
a regular network format. Implementing such rule shall
density mass points even if the areas are almost horizontal, a
fact which may be useful when it comes to some types of
surface interpolation methods.
Surface modeling factor rule can be defined as follows:
Sd = Md / sin alfa (1)
Hd = Md / tan alfa (2)
Where Sd = Maximum slope distance between successive.
Points
Md = Surface modeling Factor represented by the
maximum height difference between successive
points.
Alfa = slope angle between the successive points.
Hd = horizontal distance between successive
points.
Supposing that the suggested grid distance is Gd, the
horizontal distance between successive measured points shall
be Hd where the slope is steep, but shall not exceed Gd for
less sloppy areas. If these rules are implemented then a
certain quality can be reached. For example if I require my
data to be measured with a grid distance of 10m and my
surface modeling factor to be 1 meter, the I shall assure that
my model dose not deviate from the real surface by an error
grater than 50 cm which is the maximum expected
interpolating error between successive points taken as half
the height difference value between the successive points.
Data collected Using existing Maps: When digitizing
existing maps, we are to consider some more errors
depending on the digitizing method, the scale of the map, the
interval of the contours, and the validity of the data, so that
an assessment of quality can be made. This method shall
obviously present an accumulation of errors, and
overshooting contours shall create an error depending on the
steepness of that particular segment.
Interpolation and processing methods: A very effective
part of quality aspect of the surface data is interpolation, and
that is because we are seeking to represent the infinitely
continuous surface using the limited collected surface
information, so the interpolation method is very hectic and
depends on how many neighbor points it considers and with
what kind of weighting rules, the distance to the nearest
points, the density of the data, the roughness of the terrain,
and others such as surface fitting.
Digital Format or Structure of the Data: The digital
format on the other hand shows the final result, it makes a
big difference if data is collected with a certain accuracy and
represented with a less accurate method or visa versa, both
situations shall mislead the operator in a way or other, and
some digital structure formats are more convenient and
legible to be used for representing the surface fused with
planimetric data. For example contours can give a pleasant
looking when superimposed on planimetric data to express
the surface, although TIN data are far more accurate, raster
data on the other hand provide a general impression about the
terrain.
Then choosing between the different Surface structures shall
dominantly depend on the purpose the data should be used
for, it can be one type or a combination of several structures.
and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
70
Total accuracy of data: If we suppose that the total
deviation from the truth height at a certain point on the
surface is equal to the sum of deviations resulted due to the
accumulation of errors, then
dz = Sn + Sm + Si + So (3)
Where dz = total deviation of the z value of a surface point
Sn = error due to the control network used
Sm = error due to the measurement method
Si = error due to interpolation
So = other accumulation errors
Thus when planning data collection methods and modeling
aspects the sum of these errors has to be taken into
consideration and the result should be reflected with the data
to overcome problems accruing due to misleading visual
representations of the terrain models.
Other errors can be such as the incompleteness of measured
break lines.
2-2 Representing Quality of the data
Among other aspects the quality and validity information
should be encountered with the data. Some proposed
methods are for example creating a buffer around contours in
a separate layer showing the maximum deviation of the
values of the contours, creating error ellipses around
measured points, and registering meta data of the feature
classes showing the validity of the data such as date of
measurement, source of data, assessed accuracy,
interpolation method, etc.
3. Modeling Rules
Rules should be set for height data in a geodatabase, in such
a manner as topology rules, implementing topology rules for
parcels as an example such as parcels must not overlap, shall
insure a certain quality of the planimetric data, a rule that
data can be validated for. If we design similar useful rules for
height data we shall be able to insure a certain quality and
cleanness.
A simple rule is not to allow contours to cross, may show the
effected blunders created by certain DTM software when
interpolating and smoothing of the contours.
A rule setting the maximum surface modeling factor Md as
in equation (1) to be 2m, shall validate all model points to be
within this range.
Setting a rule for changing the contour interval as a function
of the displaying scale and average steepness of the zoomed
area, shall insure a pleasant representation without covering
the planimetric details under the data .
Other rules which can filter TIN structures, such as
maximum triangle side length.
Considering such modeling rules and others such as color
and thickness of contours in certain areas depending on the
background, shall prevent many errors that can not be seen
other wise without checking the whole data against those
rule, it shall provide a certain limit of automation, and
comicality.
Intei
Con
The
mea
misl
heig
mea
or a
befo
Enc
user
rule:
such
strug
Ref:
Dav
Repi
Ace
Bris
Detr
Peri
84
Mar
Mod
No.
Step
ICA
Abd
Exp:
Abd
And
Abd
Qua
Engi
Detr
Hun
Detr
Syst
Ack
Here
Dert
all i
throi
