DB-SMoT: A direction-based spatio-temporal clustering method

—Existing works for semantic trajectory data analysis have focused on the intersection of trajectories with application important geographic information and the use of the speed to find interesting places. In this paper we present a novel approach to find interesting places in trajectories, considering the variation of the direction as the main aspect. The proposed approach has been validated with real trajectory data associated to oceanic fishing vessels, with the objective to automatically find the real places where vessels develop fishing activities. Results have demonstrated that the method is very appropriate for applications in which the direction variation plays the essential role.

• Type: Journal
• Year: 2010
• Where: IS
• Authors: Jose Antonio M. R. Rocha, Valéria Cesário Times, Gabriel Oliveira, Luis Otávio Alvares, Vania Bogorny
• Keywords: Artificial Intelligence, Interesting Places, IS 2010, Semantic Trajectory Data, Trajectory Data