Using Deep Learning to Count Albatrosses from Space

Bowler, Ellen, Fretwell, Peter, French, Geoffrey and Mackiewicz, Michal ORCID: (2019) Using Deep Learning to Count Albatrosses from Space. In: IEEE International Geoscience and Remote Sensing Symposium, 2019-07-28 - 2019-08-02.

[thumbnail of Accepted_Manuscript]
PDF (Accepted_Manuscript) - Accepted Version
Download (2MB) | Preview


In this paper we test the use of a deep learning approach to automatically count Wandering Albatrosses in Very High Resolution (VHR) satellite imagery. We use a dataset of manually labelled imagery provided by the British Antarctic Survey to train and develop our methods. We employ a U-Net architecture, designed for image segmentation, to simultaneously classify and localise potential albatrosses. We aid training with the use of the Focal Loss criterion, to deal with extreme class imbalance in the dataset. Initial results achieve peak precision and recall values of approximately 80%. Finally we assess the model’s performance in relation to interobserver variation, by comparing errors against an image labelled by multiple observers. We conclude model accuracy falls within the range of human counters. We hope that the methods will streamline the analysis of VHR satellite images, enabling more frequent monitoring of a species which is of high conservation concern.

Item Type: Conference or Workshop Item (Paper)
Faculty \ School: Faculty of Science > School of Computing Sciences
UEA Research Groups: Faculty of Science > Research Groups > Colour and Imaging Lab
Depositing User: LivePure Connector
Date Deposited: 24 Jun 2019 14:30
Last Modified: 22 Mar 2024 00:36
DOI: 10.1109/IGARSS.2019.8898079


Downloads per month over past year

Actions (login required)

View Item View Item