The integration of various data sets can be the answer for geoscientific questions on the one hand, but a disadvantage on the other hand, due to the differences in representation and content. Although geoscientific data sets typically refer to the same physical data source – the earth surface – and therefore also relate to topographic objects, these data sets differ in geometry, accuracy and actuality in most cases. In former times differences between analogue maps were not as apparent as today when different data sets are overlaid in a modern GIS-application. Integrating different data sets – in our case topographic data and geoscientific data – allows for a consistent representation and thus for the propagation of updates from one data set to the other. This problem leads to three steps, namely harmonisation, change detection and updating which are necessary to ensure consistency, but hardly practicable when performed manually. For a harmonization of data sets of different origin, firstly the revelation of semantic differences is required; to this end, the object catalogues are compared and semantically corresponding objects are identified. In this step, also the cardinality of possible matchings between the objects in the different representations is determined (1:1, 1:n, n:m). The identification of geometric differences between the one-layered geoscientific and the multi-layered German digital map (ATKIS) will be fulfilled in the next step. In order to identify corresponding object-pairs between the data sets, different criteria like area, shape and position are used. Due to different levels of generalisation the detection of matches between groups of objects and single objects is implemented. Corresponding objects which have been selected through semantic and geometric integration are investigated for change detection using intersection methods. The geometric differences which are visible as discrepancies in position, scale and size due to simple superimposition will lead to unsatisfying results. Therefore, the iterative closest point (ICP) algorithm is implemented to achieve the best fit between the objects. The evaluated results can be classified into three types, of which two types can be handled automatically, and for one type an automatic proposal is given by the software. This leads to a significant reduction of time and resources because the approach reduces the objects to be investigated manually to only those situations where manual intervention is inescapable. The paper gives an overview of the problem and focuses on the geometric integration, especially on the matching of groups of objects and the adaptation of the object’s shape.
[1]
Paul J. Besl,et al.
A Method for Registration of 3-D Shapes
,
1992,
IEEE Trans. Pattern Anal. Mach. Intell..
[2]
T. Badard.
On the automatic retrieval of updates in geographic databases based on geographic data matching tools
,
1999
.
[3]
Martien Molenaar,et al.
Ontology-Based Geographic Data Set Integration
,
1999,
Spatio-Temporal Database Management.
[4]
Soe-tsyr Yuan,et al.
Development of Conflation Components
,
1999
.
[5]
Volker Walter,et al.
Matching spatial data sets: a statistical approach
,
1999,
Int. J. Geogr. Inf. Sci..
[6]
A. Dam.
OBJECT RECOGNITION AND SCENE CLASSIFICATION FROM MULTISPECTRAL AND MULTISENSOR PIXELS
,
1999
.
[7]
C. Heipke,et al.
MODELLING THE INTEGRATION OF HETEROGENEOUS VECTOR DATA AND AERIAL IMAGERY
,
2003
.
[8]
Monika Sester,et al.
SEMANTIC AND GEOMETRIC INTEGRATION OF GEOSCIENTIFIC DATA SETS WITH ATKIS – APPLIED TO GEO-OBJECTS FROM GEOLOGY AND SOIL SCIENCE
,
2003
.
[9]
K.-H. Anders,et al.
A Hierarchical Graph-Clustering Approach to find Groups of Objects
,
2003
.