Line Simplifcation Using the Hausdorff Distance as Error Metric

Please enable javascript in your browser.

Naseri, Shahlla | 2014

940 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 47910 (02)
University: Sharif University of Technology
Department: Mathematical Sciences
Advisor(s): Zarei, Alireza
Abstract:
Due to recent advancements and wide usage of location detection devices, huge amount of data are collected by GPS and cellular technologies which exhibits moving objects trajectories. Using this information, it is possible to track a set of objects over a long period of time, as happens for instance in studying animal migration. Always, in these situations it is undesirable or even impossible (due to process and storage limits) to store the complete stream of positioning data. Then, simplifying a trajectory as a data reduction technique is the option for resolving such problems. Moreover, there is an increasing interest in queries capturing ”aggregate” behavior of trajectories as groups like identify groups of suspicious people, resource allocation, security management and infectious disease. In this thesis, we first proposed a method for constructing a band with specified edges in which trajectories of moving objects lie. As the second part of the thesis, an algorithm is proposed for dividing the initial band into several sub-bands each of which contains trajectories of ”similar” objects
Keywords:
Line Simplification ; Computational Geometry ; Approximate Algorithm ; Data Flow