An Intelligent Data Analysis Framework for Supporting Perception of Geospatial Phenomena

Abstract

Land use and urban development surveys involve the interpretation of a large volume of data coming from satellite images processing as well as from remote sensors networks. In order to facilitate this interpretation, the development of a multipurpose Intelligent Data Analysis (IDA) framework for supporting geographical data perception is proposed here. The framework makes use of semantic technologies and relies on a novel knowledge model composed by a foundational ontology (DOLCE Ultra-Lite, also called DUL), three core reference ontologies (the Temporal Abstraction Ontology or TAO, the Semantic Sensor Network ontology or SSN and the SWRL Temporal Ontology or SWRLTO) and two specific domain ontologies (the Urban Ontology or URO and the Geographic Data ontology or GeoD, developed by our team). They play different and well specific roles in the whole process of perception. The paper shows how to apply SSN to manage measurements of geographical regions provided by satellite images processing software. In a similar way, TAO has been extended to deal with the abstractions resulting from geographical data interpretation. An example shows a SWRL based implementation of a perception process that gradually abstracts geographical features and objects.