Scientific profile:

Mariana Belgiu graduated as an MSc in Territorial Planning and Sustainable Development at the West University of Timisoara, Romania (2007) and MSc in Applied Geoinformatics at the University of Salzburg, Austria (2009). From July 2009 to April 2010 she worked as a research associate at the Institute for Geographic Information Science, Austrian Academy of Sciences and at the Centre for Geoinformatics (Z_GIS), Salzburg Austria. She was involved in several Spatial Data Infrastructure related projects in Austria and Central Asia. Currently, since April 2010, she is a PhD Student at the Institute for Geographic Information Science. Her main research interests are: ontology, Spatial Data Infrastructure, data harmonization and semantics within GIScience.

Education

2001 - 2005  B.A Geography- English, West University of Timisoara

2005 - 2007  MSc Territorial Planning and Sustainable Development, West University of Timisoara

2007 - 2009  MSc Applied Geoinformatics, University of Salzburg

2010 -          PhD Student, University of Salzburg

2011 -          Associated PhD Student, GIScience Doctoral College, Salzburg, Austria

Professional career 

2005 - Invited Teaching Assistant, West University of Timisoara (winter semester 2005-2006 - geomorphology lab)

2009 - Research Associate at the Austrian Academy of Sciences, Institute for Geographic Information Science (July 2009 - December 2009)

2010 - Research Associate, Centre for Geoinformatics (Z_GIS), University of Salzburg (January 2010 - March 2010)

2010 - PhD Student, Austrian Academy of Sciences (ÖAW), Institute for Geographic Information Science

Research Cluster: Time and Process; Spatialization, Media and Societa

PhD Thesis Topic: Prototype Entities for Objects to be extracted from satellite imagery - Land Cover Scenario

Abstract:

In the last years, remotely sensed data has developed tremendously in terms of accuracy, spatial, temporal and spectral resolution and resulting volume. In order to efficiently use such large data volume, there is an increasing need for optimizing the methods for automated information extraction. To extract information from satellite imagery, analysts have to translate a priori knowledge (definition of object of interest, its appearance in image and image processing knowledge) into classification rules. Unfortunately, knowledge used to extract information from image is not always structured in a systematic, objective and transparent way. In the absence of an explicit specification of this knowledge, the process of information extraction remains subjective, time consuming, error-prone and hardly reproducible. In our research we make use of the current progress in ontology engineering to formalize domain knowledge and the appearance of defined concepts in remote sensing imagery. The overall goal of our approach is to assess the versatility and transferability of object categorization developed on ontological bases. Land cover domain serves as a running scenario across this research. The developed image analysis framework is applied to WorldView-2 satellite imagery.

Publications:

Research Areas: SDI, semantics, metadata, Remote Sensing