Morphing techniques for creating and representing spatiotemporal data in GIS

Session D1: GIS and Environmental Modeling for Decision Support

With the development of remote sensing technologies allowing capturing and transmitting geo-referenced data repeatedly along time, there are many applications demanding for efficient tools to deal with spatiotemporal data. However database research on moving objects with extent has mainly focused on spatiotemporal data models and query languages leaving several issues to be solved regarding, for example, the acquisition of spatiotemporal data. This work deals with the application of 2D polygonal morphing techniques to create spatiotemporal data representations of moving objects that may change location, size or shape continuously over time. The aim is to transform a sequence of observations representing a moving object at different time instants into a continuous movement representation suitable to be loaded into spatiotemporal databases. This work also investigates several strategies to minimize users' intervention during the processing of a sequence of observations and presents an evaluation of the reliability of movement representations using real data. The movement of icebergs in the Antarctic seas is used as case study and the data sources are sequences of satellite images capturing the position and shape of the icebergs at different dates.