Integration of the segments from analysis of a satellite image into an Object Oriented Ontology

In recent years the field of computer science has grown in an almost exponential rate; this growth has changed the world. Thanks to this growth we may use technology to remove the human limitations of processing visual information, this is known as remote sensing. LSIIT laboratory has researched many algorithms for remote sensing. LSIIT laboratory has also developed a Java based desktop application called “Mustic”. Mustic uses a collection of machine learning algorithms to process satellite images and output a file in Attribute-Relation File Format (ARFF). LSIIT laboratory recently set as a goal to take information outputted from Mustic and classify the segments of information from the output into geographical objects. To do this LSIIT laboratory has developed a object oriented ontology that describes the logic that human geographers use to classify geographical objects on a map. This ontology was created on Protégé which is a Java base ontology creation software. My work consisted of designing rules that would allow there to be a translation between the values that Mustic outputs and the values that the Protégé ontology uses to classify objects. We designed a set of logical statements, from these I decided to design a plug-in for Protégé that automatically creates the relationship between an ARFF file and the object oriented ontology. The result is a plug that reads, processes, saves files in ARFF, and generates geographical objects for the ontology. This research project was helpful because it added abstraction and automation to the geographical object recognition process.