)12 
NCE 
rdogan, 
tions and 
jlete data 
countable 
bases and 
costs and 
different 
metadata, 
producers 
ling, data 
eral rules 
hold, data 
aken into 
checking 
isers. But 
11 method 
pplication 
ealization 
discusses 
'oduction. 
ic Treaty 
both data 
into three 
of quality 
at can be 
y control, 
ty control 
1ecks that 
ly, these 
at allow 
required. 
he terrain 
mapping. 
mentation 
odata and 
cts to get 
terms of 
is a must. 
make sure 
:d quality. 
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 
Such a program has to cover a cycle of procedures for quality 
control that warranties to keep the specified quality. Each 
phase in quality control assessments defines one aspect of 
quality requirements for the underlying information model. 
Thus, quality management should be the initial phase in data 
processing, data analysis, maintenance or homogenization of 
different data sets to make sure a well-defined result in any of 
these processes. (Busch and Willrich, 2002) 
Many studies are performed about the quality of geodata. In 
one of these studies, a framework to support quality driven 
large-scale geospatial data integration (QGM) was described 
by Thakkar et al (2007). The key contributions of their 
framework are: (1) the ability to automatically estimate quality 
of data provided by a source by using the information from a 
source of known quality, (2) declarative representation of both 
the content and the quality of geospatial data provided by 
sources, and (3) a quality-driven query answering technique for 
geospatial data. Their experimental evaluation using over 1200 
real-world sources show that QGM not only provides better 
quality data compared to the traditional data integration 
systems, it also has lower response time. 
Spatial data quality is well-known by academia and industry 
but usually in different context. The research on spatial data 
quality stated several issues having practical use such as 
descriptive information as metadata, fulfilment of spatial 
relationships among data, integrity measures, geometric 
constraints etc. The industry and data producers realize them in 
three stages; pre-, co- and post data capturing. The pre-data 
capturing stage covers semantic modelling, data definition, 
cataloguing, modelling, data dictionary and schema creation 
processes. The co-data capturing stage covers general rules of 
spatial relationships, data and model specific rules such as 
topology and model building relationships, geometric 
threshold, data extraction guidelines, object-object, object- 
belonging class, object-non-belonging class, class-class 
relationships to be taken into account during data capturing. 
And post-data capturing stage covers specified QC benchmarks 
and checking compliance to general and specific rules. 
Vector data represents one major category of data managed by 
GIS. Based on geo-spatial data standards and integrity rules 
GIS vendors and data producers build QC and QA guidelines 
and apply them in production workflow. The vector data 
quality criteria are different from the views of producers and 
users. But these criteria are generally driven by the needs, 
expectations and feedbacks of the users. 
This paper presents a practical method which closes the gap 
between theory and practice. Development of spatial data 
quality concepts into developments and application requires 
existence of conceptual, logical and most importantly physical 
existence of data model, rules and knowledge of realization in 
a form of geo-spatial data. The applicable metrics and 
thresholds are determined on this concrete base. 
This study discusses application of geo-spatial data quality 
issues and QC procedures in the topographic data production. 
Firstly we introduce MGCP data profile of NATO DFDD, the 
requirements of data owner, view of data producers for both 
data capturing and QC and finally quality assurance to fulfil 
user needs. Then, our practical and new approach which 
divides the quality into three phases is introduced. Finally, 
23 
implementation of our approach to accomplish metrics, 
measures and thresholds of quality definitions is discussed. In 
this paper, especially geometry and semantics quality and 
quality control procedures that can be performed by the 
producers are discussed. 
2. QUALITY ASSURANCE OF VECTOR DATA 
2.1 What is Vector Data? 
Vector data provide a way to characterize real world features 
within the GIS environment. A feature is anything you can see 
on the landscape. Think about you are at a high position on the 
field. When you look at down, you can see many features like 
forests, houses, roads, trees, rivers etc. Each of these things 
would be a feature when we characterize them in a GIS 
program. Vector features have attributes, which consist of text 
or numerical information that describe the features. 
A vector feature has its shape represented using a special 
geometry. The geometry is made up of one or more 
interconnected vertices. A vertex describes a position in space 
using a x, y and optionally z axis. Geometries which have 
vertices with a z axis are often referred to as 2.5D since they 
describe height or depth at each vertex, but not both. When a 
feature's geometry consists of only a single vertex, it is referred 
to as a point feature. Where the geometry consists of two or 
more vertices and the first and last vertex are not equal, a 
polyline feature is formed. Where four or more vertices are 
present, and the last vertex is equal to the first, an enclosed 
polygon feature is formed. (www1, 2011) 
2.2 Quality of Vector Data 
The quality elements described in ISO (International 
Organization for Standardization) 19113 are completeness, 
logical consistency, positional accuracy, temporal accuracy, 
thematic accuracy. Also they can be defined as measures of 
quality such as closeness to the actual value, spatial accuracy 
of position, shape, size/area of features, "currentness" of data 
and completeness of attribute values. Another side of quality 
should be thought as absence of contradictions in the data and 
conformance of the data to rules. 
Also there are other classifications made for the quality of 
vector data. In one of these classifications made by Subbiah et 
a. (2007), quality parameters are defined as accuracy, 
resolution, completeness, and types. Accuracy of geospatial 
data is defined in terms of tuple (attribute, value), where 
attribute refers to a geographic concept/object and the value is 
its measurement. It is assumed that geospatial service 
providers provide data that conform to such tuples and that 
there is an objective assessment of all concept values. 
Resolution refers to the amount of detail that can be 
determined in space, time or theme. Vector data can be 
represented in either fine or coarse density. Scale of maps can 
be thought as resolution. The coarser the data is, the less 
information is available about vector points of an object’s 
shape. Resolution is also related to accuracy because the level 
of resolution affects the database specification against which 
accuracy is assessed. Completeness refers to the absence of 
omissions in a provider database. Completeness is distinct 
from accuracy in that the errors that result in lack of 
completeness are not incorrect encoding of object values.