International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B4, 2012 
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia 
Instead, when a service provider fails to keep its database 
updated with latest data, it is considered to have incomplete 
data. Data types refer to the format of desired data. Even 
though the area of geospatial data interoperability has made a 
lot of progress, various reasons still exist that lead clients to 
request specific type of data format. 
Nowadays term of “level” is used to define the resolution of 
vector data. VMap (vector-map) is a good example for this. 
VMap Level 0 corresponds to approximately 1:1.000.000 scale 
and Level 1 to 1:250.000. Especially globally produced vector 
data is labelled with levels. A good example to such kind of 
global production is MGCP. In that program level-2 data that 
corresponds to 1:100.000 or 1:50.000 scale is produced by 
different nations. The project represents the most current 
evolution of a 10-year, global VMap Level 1 effort that began 
in 1993 and was revamped in 2003. In such kind of projects, 
quality becomes more important, mainly to prevent the 
production of non-harmonized data between the participant 
countries. 
Vector data needs a lot of work and maintenance to ensure that 
it is accurate and reliable. Inaccurate vector data can occur 
when the instruments used to capture the data are not properly 
set up, when the people capturing the data aren't being careful, 
when time or money don't allow for enough detail in the 
collection process, and so on. If you have poor quality vector 
data, you can often detect some of these quality lacks when 
viewing the data in a GIS. (wwwl, 2011) 
The usefulness of the quality measures depends on the 
application. It is not always clear to decide on how many 
quality parameters can be introduced to describe the quality of 
data. The number of quality parameters can be very large 
because quality varies spatially and temporally. Defining the 
quality measures is already a very actual topic in the 
standardization process. (Ragia, 2000) 
Level of the quality should be adjusted carefully. Quality and 
efficiency or productivity are conflicting or opposite aspects. If 
the quality of the product is selected very high, it increases the 
quality control period and decrease the productivity. The 
relation between these objects can be seen as in Figure 1. So 
the quality, time for quality control and productivity should be 
optimized according to the needs. 
Q 
@ me = 
  
7 > 
QC Process 
Productivity 
  
Figure 1. Quality vs QC Process and productivity relation 
2.3 Quality Assurance of Vector Data 
For rating the quality of geodata certain set of measures are 
needed, which give us expressive, comprehensive and useful 
criteria. À coarse subdivision of quality measures into two 
categories can be done, which due to the following arguments 
are important for practical applications: 
1. Quality measures that concern consistency with the data 
model, 
2. Quality measures that concern consistency of data and 
reality within the scope of the model. 
A complete check of the first category can be performed 
automatically within a database or GIS without any additional 
data. This inspection can be done exhaustively, i.e. the whole 
area covered by the data can be checked. On the other hand the 
comparison of data and reality is much more expensive. 
Performing it for the whole area requires much more effort. 
(Busch, 2002) 
In this paper quality control procedures, which are used for 
MGCP production in General Command of Mapping, to 
accomplish the certain quality measures and the experiences 
from these procedures are discussed. Our quality control 
procedures consist of parameters specifying the following 
quality aspects: topology, geometry, completeness of features 
and attributes, logical consistency. To succeed these aspects 
some automatic, semi automatic or manual quality checks are 
used. 
The data quality is assured in feature and feature-class levels. 
The error inspection procedure comprises consistency, 
completeness and correctness control categories. QA software 
is the main QC tool for consistency checking and topology. 
This software checks the topology errors, some geometry errors 
like connection, overlaps etc, attribution conformity and 
compatibility errors. Some of the found errors are corrected 
later automatically or manually but without checking the error. 
Some of the found errors should be checked over the data and 
other sources since they can be false positives, which means 
that they seem as errors but in fact they are not errors. The 
completeness of geometry and attributes is tried to be 
guaranteed based on four levels control approach. 
At the first level, vector data is controlled over the source 
imagery and other ancillary sources. At this stage the captured 
data is controlled according to the technical prerequisites. A 
control stuff checks all captured data and looks at; 1) if the 
features on the topography captured as the correct features in 
feature and attribute dictionary, 2) if the correct attributes are 
assigned to the captured features, 3) if the feature captured 
with the correct geometry according to its size (point, line or 
area), 4) if the features corrected with the needed density, 5) if 
the features captured with the needed geometric location 
accuracy. In Figure 2, a dry river found in this control, that is 
not captured but should be captured according to the defined 
standards, is shown. This control takes approximately 95 to 10 
of the production time of data capture and approximately 965 
additional conditions of the captured data are detected. 
24 
  
  
TI 
pr