News and events
13 April 2011
Autonomous Monitoring of Nesting Seabirds Using Computer Vision
Presented By Dr Patrick Dickinson & Dr Chunmei Qing, University of Lincoln
- Room N225, Haslegrave
About this event
Abstract: Seabird populations are an important and accessible indicator of the health of marine environments: variations have been linked with climate change, pollution levels, and changes in fish-stock levels. Manual monitoring is currently used to monitor the development of important colonies, but is necessarily labour-intensive and error-prone. The use of computer vision based surveillance offers the potential to automate the collection of population-level data on a scale not currently possible using manual methods. We are working with the Centre for Computational Ecology and Environmental Science group at Microsoft Research Cambridge on developing a vision-based system to monitor a population of Common Guillemots nesting on Skomer Island, West Wales. Our work focuses on the robust detection and localisation of Guillemots in high-definition images and videos of cliff nesting areas. The challenges in this work are highly contextualised, and include camouflage of birds, crowding and image quality. The approaches we have developed include a region-based background model used with a MRF pixel classifier, and a feature-based head and body detector based on a combination of Histograms of Oriented Gradients (HOG) and Local Binary Pattern (LBP) features, to capture characteristic local edge/shape and texture signatures. We present our feature-based detector in details, and discuss ongoing and future research directions, and applications.