GPS location history data mining and anomalous detection: the scenario of bar-headed geese migration

Abstract

It is important to discover common movement sequences and uncommon behaviors during the migration of wild birds. In this paper, we propose a new approach to analyze the GPS location history data of migratory birds. The stopover sites are first extracted from the location history data of birds, and their movement sequences are generated automatically. Then, a consistency calculation method is introduced for calculating the movement sequence consistency degrees among the birds. The common movement sequences and uncommon behaviors can be recognized on the basis of consistency. We conducted experiments on the data collected from bar-headed geese captured in the Qinghai Lake region. The experiment results indicate the correctness of our approach.