Monika Sester

Linking Objects of Different Spatial Data Sets by Integration and Aggregation

In order to solve spatial analysis problems, nowadays a huge amount of digital data sets can be accessed: cadastral, topographic, geologic, and environmental data, in addition to all kinds of other types of thematic information. In order to fully exploit and combine the advantages of each data set, they have to be integrated. This integration has to be established at an object level leading to a multiple representation scheme. Depending on the type of data sets involved, it can be achieved using different techniques.Such a linking has many benefits. First, it helps to limit redundancies and inconsistencies. Furthermore, it helps to take advantage of the characteristics of more than one data set and therefore greatly supports complex analysis processes. Also, it opens the way to integrated data and knowledge processing using whatever information and processes are available in a comprehensive manner. This is an issue currently addressed under the heading of ‘interoperability’.Linking has basically two aspects: on the one hand, the links characterize the correspondence between individual objects in two representations. On the other hand, the links also can carry information about the differences between the data sets and therefore have a procedural component, allowing the generation of a new data set based on given information (i.e., database generalization).In the paper three approaches for the linking of objects in different spatial data sets are described. The first defines the linking as a matching problem and aims at finding a correspondence between two data sets of similar scale. The two other approaches focus on the derivation of one representation from the other one, leading to an automatic generation of new digital data sets of lower resolution. All the approaches rely on methodologies and techniques from artificial intelligence, namely knowledge representation and processing, search procedures, and machine learning.

Visualization in an early stage of the problem-solving process in GIS

Abstract Methods of user–computer interaction have remained largely unchanged since the introduction of graphical user interfaces and their popularization by the Apple Macintosh in the early 1980s. Most of todays applications rely on primitive modalities, such as typing and pointing for input generation, which works well for a host of common business applications, but falls short for more complex tasks. To improve the interaction between user and computer we propose a concept that allows people to visualize their ideas, problems, or instructions during the initial phase of an interaction with a computer by augmenting traditional interaction modalities with sketching, gesturing and talking. This approach leads to a more natural user–computer interaction and enhances a users ability to find solutions to a problem. We suggest that computers become actively involved in the process of problem formulation and that they provide support and give advice where this is adequate. This leads to a process of incremental problem formulation where user and computer are able to better visualize the actual task and fewer misunderstandings occur. Geographic information systems (GIS) would benefit from improved user interaction techniques. GIS are inherently complex and an interaction is often tedious, mostly because such systems are based on sequential and nonspatial input methods that lack the capability of expressing spatial concepts appropriately. We advocate for a visualization in an early stage of the problem solving process in GIS and discuss its advantages and challenges. The paper gives application examples and discusses future research topics.

Object Location Based on Uncertain Models

The paper describes a concept for object location, when not only image features but also the model description is uncertain. It contains a method for probabilistic clustering, robust estimation and a measure for evaluating both, inaccurate and missing image features. The location of topographic control points in digitized aerial images demonstrates the feasibility of the procedure and the usefulness of the evaluation criteria.

Test on image understanding

Working Group III/3 of ISPRS, entitled `Semantic Modelling and Image Understanding, has organized a test on image understanding. The basic goal of this test is the integration of different information sources in the image interpretation process. In addition to aerial images all sorts of additional information can be used to derive a more reliable result, e.g. color images, stereo, GIS-planimetry. The test deals with man-made objects, for instance houses, streets, or fields of land, which are to be detected and reconstructed from the image data and further information.