296 
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B6b. Beijing 2008 
The selected conventional, standardized GIS client /server 
interfaces like WMS and WFS can also be applied for mobile 
services. On basis of the available mobile communication 
technologies like WLAN, GPRS, UMTS and Bluetooth, it is 
possible to network different mobile system components. It is 
conceivable that the bandwidth of UMTS and WLAN supports 
the transfer of larger amounts of data. So the problem of the 
small bandwidth faced by previous technologies is solved and 
the principal requirements for online mobile access to 
heterogeneous databases are meeting. The usage of standardized 
interfaces and therewith the avoidance of proprietary 
developments leads to an open structure of the GIS platform. 
4. FIELD DATA COLLECTION AND QUALITY 
ASSURANCE 
Mobile data acquisition provides a new chance for the fieldwork 
in geosciences. A mobile data acquisition system has been built 
up based on OGC’s Transactional Web Feature Service (WFS-T) 
and extensions. Therewith using this system mobile field 
workers cannot only have interoperable access to heterogeneous 
databases in the field, but also use transaction operations like 
inserting, deleting and updating data. The acquisition system is 
generic and can be easily adapted to different geosciences 
applications. The direct transfer of newly collected data into 
databases requires data quality assurance directly in the field 
[13] . 
Mobile data acquisition enables the inclusion of functions that 
support the improvement of data quality during outdoor data 
collection. One approach to support field workers during data 
collection is the ontology-based capture of data, which is based 
on a list of rules that can provide diagnosis for the user to make 
decisions. These plausibility checks enable the user to correct 
errors in the field, thus guaranteeing a higher quality of data 
[14] . As parts of the data model, spatial, topological and 
semantic quality constraints and combination of them are 
defined to express the plausibility rules. Besides those 
constraints, the instruction information about how users should 
react to the possible errors is also defined. 
4.1 Ontology based quality constraints definition 
During mobile data acquisition, data quality plays an important 
role since all of the collected data in the field are supposed to be 
immediately transferred to the databases. In order to ensure data 
quality and reduce error risks, a quality assurance method 
integrated into the mobile acquisition workflow has to be taken 
into account. 
In this paper, constraints have been composed with the mobile 
data acquisition system as an extension to the data application 
schema. In that conjunction OGC’s WFS-T is used to provide 
the mobile system with the application schema, moreover the 
transactional services of WFS-T can be used to transfer newly 
collected features to a remote server. For such task it is well 
known that the WFS-T server is able to provide an XML 
schema document according to the feature types listed in the 
request. This XML schema document is a GML application 
schema that can be used to validate collected feature instances 
which should be sent to the remote server. However the 
application schema only includes information about the features, 
e.g. geometry type, attribute name, attribute data value type of 
and etc. With this information the data can only be assured with 
regard to geometry type, attribute fields and data types. But 
from the users point of view there are more requirements 
especially with regard to data integrity which should also be 
checked in the field. For example, a topological constraint like 
“a hiking way is not allowed to be intersected with a ditch” 
cannot be provided by the normal GML application schemas. 
Therefore a way of extending the GML application schema has 
to be investigated. An ontology based method can be used for 
that. The selection of a certain ontology formalization language 
for the definition of quality constraints depends primarily on its 
expressive power. SWRL member submission document of 
W3C [15], which bases on the combination of sublanguages of 
Web Ontology Language (OWL) and Rule Markup Language 
(RuleML) gives a new chance for the definition of logical 
relationships in an ontology language. Therefore, the Semantic 
Web Rule Language (SWRL) is attempted for defining data 
integrity constraints. Topological and semantic constraints as 
well as their combinations are defined in the ontology language 
to ensure the data quality. For the definition of such constraints, 
spatial relations like “intersect” or “disjoint” have to be used. 
Therefore a more detailed description is given in [16]. In SWRL 
a rule axiom consists of an antecedent (rulemkbody) and an 
consequent (rulemkhead). Informally a rule may be understood 
by the meaning that if the antecedent holds (is “true”), then the 
consequent must also hold. An example to how the concept is 
implemented in SWRL is given in Figure 3. 
The annotations in line 2 to line 5 of the example enable for a 
further description of the constraint. The “constraintID” item 
contains the index of this constraint. In our definition “severity” 
value can have three different values: “strict”, “avoid violation” 
and “apply with caution, user’s reaction necessary”. The first 
one means that a violation of the constraint is illegal and the 
violating data has to be changed. The two other values leave it 
up to the user’s decision with respect to what has to be done in 
case of a violation. Therewith it is possible to use constraints as 
a description of (maybe unusual) relations of objects, which are 
not strictly forbidden but nevertheless have to be checked. The 
third value additionally requires some reaction by the user, e.g. 
the user should record the current situation according to the real 
world environment for the other possible users. The “comment” 
and the “correctionlnstruction” items provide users with helpful 
information about how to react to the violation. 
Line 6 to line 23 contain the antecedent part that shows the 
assignment of two variables way and ditch as Way and Ditch 
objects, and the Boolean data type attribute 
“publiclyAccessible” (which means whether the way is closed 
for public access or not depending on its availability for walking 
on) of the way is true. Line 24 to line 31 presents the consequent 
part that defines a relation between these two spatial objects. 
The “dWithin” item is a spatial relation which means two 
spatial objects disjoint with each other within a certain value. 
Because of the ontology based quality constraints are also based 
on XML structure, they can be easily attached to GML 
application schema. The constraints are encoded as annotations 
to each GML application schema with respect to their 
corresponding feature classes. Therewith the quality 
information defined in the constraints is transferable and 
available for the users during the data acquisition workflow, and 
the quality assurance task can be implemented based on that.