Supporting Distributed Geo-Processing: A Framework for Managing Multi-Accuracy Spatial Data

In the last years many countries have developed a Spatial Data Infrastructure (SDI) to manage their geographical information. Large SDIs require new effective techniques to continuously integrate spatial data coming from different sources and characterized by different quality levels. This need is recognized in the scientific literature and is known as data integration or information fusion problem. A specific aspect of spatial data integration concerns the matching and alignment of object geometries. Existing methods mainly perform the integration by simply aligning the less accurate database with the more accurate one, assuming that the latter always contains a better representation of the relevant geometries. Following this approach, spatial entities are merged together in a sub-optimal manner, causing distortions that potentially reduce the overall database quality. This thesis deals with the problem of spatial data integration in a highly-coupled SDI where members have already adhered to a common global schema, hence it focuses on the geometric integration problem assuming that some schema matching operations have already been performed. In particular, the thesis initially proposes a model for representing spatial data together with their quality characteristics, producing a multi-accuracy spatial database, then it defines a novel integration process that takes care of the different positional accuracies of the involved source databases. The main goal of such process is to preserve coherence and consistency of the integrated data and when possible enhancing its accuracy. The proposed multi-accuracy spatial data model and the related integration technique represent the basis for a framework able to support distributed geo-processing in a SDI context. The problem of implementing such long-running distributed computations is also treated from a practical perspective by evaluating the applicability of existing workflow technologies. This evaluation leads to the definition of an ideal software solution, whose characteristics are discussed in the last chapters by considering the design of the proposed integration process as a motivating example.