Abstract

In this paper we describe a proposed system for automatic visual monitoring of seabird populations. Image sequences of cliff face nesting sites are captured using time-lapse digital photography. We are developing image processing software which is designed to automatically interpret these images, determine the number of birds present, and monitor activity. We focus primarily on the the development of low-level image processing techniques to support this goal. We first describe our existing work in video processing, and show how it is suitable for this problem domain. Image samples from a particular nest site are presented, and used to describe the associated challenges. We conclude by showing how we intend to develop our work to construct a distributed system capable of simultaneously monitoring a number of sites in the same locality.