COLOR. y
Qa. 46.59
4524
14 dai
d photos
30 photo,
A report on
the report,
ailed report
en as well.
scanning is
nowadays a
f 2002, the
ap, with the
is produced
purpose of
grammetric
ends on the
is 0.5 m on
dcopies are
image, the
ical Names
ie produced
d with full
jection and
ell of 25 m
yds as a by-
product in digital orthophoto map production. Heights of the
grid are photogrammerically controlled from the year 2000. Not
all heights could be controlled as the amount of data is huge
(e.g. for one map sheet of 1 : 5000 there are more than 11.000
heights in a grid of 25 m). Thus, a proper sampling method
should be used.
Obviously, the best method for assessment of positional
accuracy is to compare the data in the model with height data of
considerably higher accuracy (e.g. with geodetic measurements
or precise GPS measurements). Due to limited finances, control
with ground data is not possible. Therefore, photogrammetric
measurements on analytical instruments are used for the
purpose of quality control.
Photogrammetric control of DEM is performed on a sample of
the area approximately 5% of map sheets in a photogrammetric
block. Usually, a photogrammetric block is one of
trigonometric sections, each composed of 50 map sheets in
scale 1 : 5000. This practically means two map sheets from one
photogammetric block, which are chosen from the SMA.
The following input data are provided for quality control:
e a list of photogrammetric blocks and map sheets that
have to be controlled,
e diafilms of stereophotos,
* DEM 25 for the controlled area generated by
producers,
e à sketch of map sheets with photo strips and photo
numbers,
e data on aerial triangulation (camera, camera
calibration certificate, ^ orientation parameters,
coordinates of tie and check points).
The requirements of the model quality are = Im in open areas.
Photogrammetric measurements for quality control are done on
precision analytical instrument. After importing the orientation
parameters of photos (from the same aerial triangulation that
was used for collecting data), the tie and control points that fall
on the particular stereo-pair are measured first. If there are no
big differences, the operator starts with measurement of heights
on the sample areas. The number of points is between 500 and
600 per map sheet. There are two categories of points: points on
open area and points in a forest. Approximately 1/3 of
controlled points are in the forest area and 2/3 of points in open
area. The position of control points is indicated in sketches
added to the report.
After photogrammetric measurements, differences on tie and
check points are computed as well as differences on controlled
points and a standard deviation (separately for open and forest
area). Two files are prepared for the report:
e complete list of control data in .mdb datafile,
® quality control report in MS Word datafile.
The report consists of technical report (general data, project
team, equipment used, summary report), table of results per
map sheet, table of differences on tie and check points, and
sketches with indicated control points positions.
3.5 Quality control of topographic data
Surveying and Mapping Authority of the Republic of Slovenia
(SMA) has started in the last few years to establish a
topographical database in large scale, which could be
141
considered also as modernized basic topographic map in scale
1: 5000 and represents a basic digital georeferenced source for
other spatial data. The model of topographic database in large
scale was first established with a prototype in 1998, the concept
was then renewed on the basis of results of user’s interview,
some pilot projects as well as experiences with new
technologies. In the framework of this initial production, around
200 map sheets were produced.
A bulk production started in 2002 and different Slovenian firms
are collecting data in accordance with the technical
specifications. The basic aim is collection of vector data of the
most demanding topographic entities with stereo-
photogrammetric means and adoption of some data from other
databases. The main users of the database are different
ministries, local communities, institutions for physical planning
and other institutions and companies (e.g. for roads, electric
network, hydrography). Topographical database in large scale
consists of different object classes which are divided into object
groups and types. The object classes are: buildings, traffic
networks, land use, hydrography and relief. The last mentioned
class is temporarily in raster form while the others are in vector
form.
Figure 3. Topographic data superimposed on orthophoto map
Buildings for the entire territory of the state were collected in
the framework of another huge project “Real estate registration
modernization project”. Geometry of buildings (shape of roofs)
is thus adopted from the Central database on buildings and in
the case of some incompleteness, this is updated in data
collection of topographic data. From data in topographic
database and additional sources (e.g. record on geographical
names, state border), standard printed outputs could be
produced (as a substitute for old analogue maps or as working
material for physical planning).
In addition to detailed technical specifications of the project,
data quality assessment is provided. First, each contractor that
collects data for topographical database is obliged to use
internal quality control procedures, which are however not
defined in advance. Second, the final results are controlled by
an authorized institution. Data control encompasses formal and
contents controls. Some phases are automated, some are visual.
