Diseño de un modelo para la representación de la dinámica espaciotemporal de objetos geográficos de tipo región mediante trayectorias semánticas en bases de datos

Abstract

The wide uses of spatial information, which goes from personal to industrial areas, has allowed to present an abstraction of the real world into a digital environment, that we use to represent, interpret, analyse and understand a great part of the object and/or the geographical phenomena that had influenced people’s lives and, in general, society. Geographic information systems were the pioneers in realizing apps for representation and analysis of spatial information with geographic characteristics, they were followed by spatial data bases as an answer to the efficient administration of that information. It has been included a temporal dimension associated to the spatial information, being this the origin of spatiotemporal database, achieving to represent not only the form and location of geographic objects, but also the time associated to that object, being able to change its position and/or shape for another instant. This representation of time is discrete in the way that the object is stored in specific instants of time. As the changes associated to the geographic objects are continuous in time, their representation on space-time data bases has the limitation of being discrete. It is considered a limitation, since the representation of the object itself doesn’t corresponds completely to reality, restricting a complete and exact interpretation of the object and/or phenomenon on time intervals. As an answer to this problem, moving objects databases emerged, which pretend to represent the shape and location changes in a continuous way, managing till the moment that representation for the moving points by change of position trajectories, generating a line of the path in time intervals. In the case of moving regions trajectories, which its representation includes shape and position changes, their generation is more complex by far, to the point that at the moment a bibliographical reference of data base of an applied model that works on a moving objects that could generate it doesn’t exist, and less use the for consulting and analysis. This doctoral thesis designs a spatiotemporal database model that generates position and shape changes trajectories of moving regions, extending the model by adding semantics to the final trajectory with the intention of characterizing and enrich that trajectory on its environmental context, and other semantic generated by the movement that is represented. As a study case for experimentation and application of the model mentioned, the geographical phenomenon of urban growth of Bogotá city is chosen, region’s trajectories are generated, these represent the urban area, and a semantic would be associated to the city growth trajectory that complement the information for enriching the crude generated trajectories. In the same way a second study case would be taken, this one will show the thaw of Santa Marta’s Sierra Nevada. The source of information will be a map sequence and satellite images in a time period, of where the information about the geographical phenomenon would be extracted by classification techniques and the results will be saved on a space-time data base, where the process for making the generation of trajectories will continue with their respective representation on the space data base. As a result of the implementation of the models proposed, is demonstrated that these models allow the representation of semantic trajectories on space data base using the study cases mentioned, it also demonstrates that consults and operations can be realized with the trajectories, in addition of their visual representation and connection with other geographical information systems.