A grid-based index and queries for large-scale geo-tagged video collections

Currently a large number of user-generated videos are produced on a daily basis. It is further increasingly common to combine videos with a variety of meta-data that increase their usefulness. In our prior work we have created a framework for integrated, sensor-rich video acquisition (with one instantiation implemented in the form of smartphone applications) which associates a continuous stream of location and direction information with the acquired videos, hence allowing them to be expressed and manipulated as spatio-temporal objects. In this study we propose a novel multi-level grid-index and a number of related query types that facilitate application access to such augmented, large-scale video repositories. Specifically our grid-index is designed to allow fast access based on a bounded radius and viewing direction --- two criteria that are important in many applications that use videos. We present performance results with a comparison to a multi-dimensional R-tree implementation and show that our approach can provide significant speed improvements of at least 30%, considering a mix of queries.