A description logic based QSTR framework for recognizing motion patterns from spatio-temporal data

Learning patterns from spatio-temporal data streams is an important problem within Artificial Intelligence. Knowledge is important for recognition of patterns. Representation of large and diverse knowledge requires formal basis. Description Logics (DLs) constitute a family of knowledge representation formalism which provide object-oriented representation with formal semantics. Qualitative spatial and temporal reasoning (QSTR) encompass efforts devoted to providing useful and well-grounded models to be used as high level qualitative descriptions of spatio-temporal change. In this paper we combine DL with QSTR and put forward a formal, explicit knowledge representation formalism for representation of motion patterns. Reasoning services of the DL system is used for recognizing motion patterns from spatio-temporal data.