N 
iv 
iv 
tv 
= 
ENT 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004 
  
McKay, 1992) has been implemented to achieve the best fitting 
between the objects from ATKIS and the geo-scientific 
elements using a rigid transformation. 
In our first approach, objects from ATKIS are considered as 
reference due to their higher geometric accuracy, and the 
objects from the geoscientific datasets are optimally fitted to 
the ATKIS objects (Goesseln & Sester, 2003). 
  
  
  
  
  
Fig. 4 : Resulting overlapping segments from mere intersection 
showing geometric differences between water 
bodies in the German digital topographic map 
(ATKIS) and in the geological map. 
At the end of the process the best fit between the objects using 
the given transformation is achieved, and a link between 
corresponding objects in the different data set is established. 
The ICP algorithm has been implemented to compensate the 
geometric discrepancies which occur due to the way the digital 
geoscientific data sets have been created using manual 
adaptation, rescaling and digitization. 
6.1 Intersection and segment evaluation 
Following these steps, intersecting objects for a proper change 
detection will lead into a more reliable result (Fig. 5) than 
simple intersection (Fig. 4). This analysis and the classification 
into different change situations is a semantic problem and will 
be conducted in close collaboration with experts from geology 
and soil science, who are also partners in the project. 
At this time of the project three different classes have been 
identified: the intersection segments can be classified according 
to their respective classifications in the original data sets in: 
e Type I! : Segment is defined as water area in both 
maps, no adaptation required, 
* Type Il : Segment in geoscientific data set has been 
any type of soil, but is defined as water-area in the 
reference data set; therefore the attribute of 
classification will be changed in the geoscientific 
map, 
oe Type III : Segment is defined as water-area in 
geoscientific data set (e.g. no soil-type definition 
available), but. no water-area in the reference data 
set. Therefore a new soil-definition is required. 
A __ù 
Type II will also be assigned to objects which are represented 
in the reference, but not the candidate data-set, this is the result 
different updating periods between the reference and the 
candidate data set, which results in outdated objects. 
While Type I and II require only geometric corrections or 
attribute adaptation and can be handled automatically, Type III 
needs more of the operators attention. 
Depending on the size and the shape of a Type III segment and 
by using a user-defined threshold, these segments can be 
filtered, removed and the remaining gap can be corrected 
automatically, this will avoid the integration of sliver polygons 
and segments which are only the results of geometric 
discrepancies and must not be taken into account. 
Different situations can cause the presence of a Type lll 
segment. Due to different natural effects like desiccation or 
man-made rerouting of a river bed, water areas have been 
changed in shape or they even disappeared from the face of the 
environment. 
After an actual topographic description is no longer available, 
there is no up to date process or method to derive a new soil 
definition automatically. As there are different ways an water 
area can disappear, there are different natural (e.g. erosion) or 
man-made (e.g. refill) processes which have influence to the 
new soil type. This new soil type could not be derived 
automatically, but there are different proposals which could be 
offered to the user by the software. An area-threshold which has 
to be defined in the near future together with the experts from 
geology and soil-science will be applied to remove Type 111 
segments which occur due to geometric discrepancies. 
As a result a visualisation will be produced showing all the 
areas where an automatically evaluation of the soil situation 
could not be derived or only a proposal could be delivered and 
manual “field work” must be performed (Fig. 5). 
The visualisation of Type III segments will already reduce the 
amount of human resources needed to detect the topographic 
changes between the geoscientific data sets and ATKIS. 
It is expected, that a high degree of automation can be achieved 
with this process. In some situations there will be an 
automatically generated suggestion from the algorithm, 
however the expertise of a human operator will still be 
mandatory in some cases in order to commit or propose another 
solution. 
7. CONCLUSION AND OUTLOOK 
The workflow presented in this paper is the result of the 
research and has been developed in close correspondence with 
the project-partners from geology and soil-science. 
The implementation of the workflow in a software protoype 
using the open source software JUMP will ensure the 
possibility of adopting the results of this project to any 
additional vector-vector integration. 
The implementation of the filtering, geometric comparison and 
the derivation of object links, together with the ICP-algorithm 
showed very good results. Processing the test data set, 
representing a standard geoscientific data sets needs less than a 
minute for water-arcas. 
At this point of the project one data set is selected as reference 
data set, which will remain unchanged while the candidate data 
sets are adjusted. If an even more accurate correspondence 
between the data sets is needed, specific geometric 
reconciliation functions for the exact adaptation of the 
geometry have to be implemented. The idea is that for that 
purpose, the individual shapes of the objects will be 
geometrically adjusted: depending on the relative accuracies of 
the original objects, an “intermediate” geometry will be 
1253