Visual Analytics of Movement by Gennady Andrienko Natalia Andrienko Peter Bak Daniel Keim & Stefan Wrobel

Author:Gennady Andrienko, Natalia Andrienko, Peter Bak, Daniel Keim & Stefan Wrobel
Language: eng
Format: epub
Publisher: Springer Berlin Heidelberg, Berlin, Heidelberg

Hence, by computing distances and filtering position records of trajectories by the distances, it is possible to detect occurrences of certain spatial relations of movers to static spatial objects or locations with particular properties (such as open spaces), specifically being inside the boundary of an object and being within a given distance from an object. These occurrences can be extracted and further analysed as spatial events.

In a similar way, interactions between two types of movers can be detected and analysed. In particular, animal ecologists are interested in interactions between predator and prey animals, such as lynxes and roe deer in our case. The algorithm for detecting encounters presented in Sect. 5.2.1 is not applicable here for two reasons. First, the algorithm does not account for different types of movers. It uses a single chronologically sorted list of position records from all trajectories disregarding the identities of the corresponding objects. Second, the algorithm applies interpolation, which would not be valid for the animal data. These data belong to the category of episodic movement data, as defined in Sect.​ 2.​9.​2, due to the large time gaps between the known positions. This means that it is impossible to detect all interactions that might have occurred between the roe deer and the lynxes in reality. However, we can hope that using only the known positions, we can find indications of at least some of these interactions.

For this purpose, we compute the spatial distances between the points from the trajectories of the roe deer and the points from the trajectories of the lynxes . The computation takes into account a temporal tolerance threshold, that is, the maximum distance in time between points from two trajectories when it is meaningful to compute the spatial distance. The temporal tolerance is necessary for dealing with data where known positions of different movers refer to diverse time moments and valid interpolation is impossible. We choose the temporal tolerance of one hour, taking into account that the data are sparse in time.

After computing the distances, we create segment filters to select only the points from the trajectories of each animal species where the distance to the nearest position of an animal of the other species is below 500 m. We consider the occurrences of such distances as indications of possible encounters between the roe deer and lynxes. From the trajectory points satisfying the filter, we create spatial events. As a result, we obtain 16 events of proximity to a lynx that occurred in the trajectories of seven different roe deer and nine events of proximity to a roe deer that occurred in the trajectory of only one lynx named Nora. Evidently, in some of the cases, the lynx was close to two or more roe deer.

To find out how many roe deer were close to the lynx at the time of each event, we use an event characterization tool, which derives new thematic attributes of events based on points from movers’ trajectories located within a given spatial distance threshold from

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.