Visual exploration of time series of remote sensing data
U. Turdukulov, Valentyn A. Tolpekin, M.J. Kraak · University of Twente Research Information · 2005
The approach described here facilitates the exploration of time-series of remote sensing data by offering a set of interactive visual functions that allows studying the behaviour of dynamic phenomena, events and evolution of phenomena over time. Essentially we simulate the human visual system by decomposing each spatial region into a number of 2D Gaussian functions. The algorithm compares the attributes of each function in successive frames and finds the continuous paths of the spatial object through time. These paths describe the characteristics of the object in each time step and for each object certain interesting ‘events’ can be described. Examples of events are: continuation, appearance/disappearance, and split/merge of the phenomena. Based on the object paths, a visualization environment with ‘temporal functionality’ is created with a wide range of tools to support interactive exploration of events and the object’s evolution. In particular, the event graph is proposed which in combination with other visualizations will make the process of detection and exploration of the dynamic phenomena independent of the perception and experience of the observer.