Details

Autor(en) / Beteiligte

• Kellenberger, Benjamin
• Marcos, Diego
• Tuia, Devis

Titel

Detecting mammals in UAV images: Best practices to address a substantially imbalanced dataset with deep learning

Ist Teil von

• Remote sensing of environment, 2018-10, Vol.216, p.139-153

Ort / Verlag

New York: Elsevier Inc

Erscheinungsjahr

2018

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• Knowledge over the number of animals in large wildlife reserves is a vital necessity for park rangers in their efforts to protect endangered species. Manual animal censuses are dangerous and expensive, hence Unmanned Aerial Vehicles (UAVs) with consumer level digital cameras are becoming a popular alternative tool to estimate livestock. Several works have been proposed that semi-automatically process UAV images to detect animals, of which some employ Convolutional Neural Networks (CNNs), a recent family of deep learning algorithms that proved very effective in object detection in large datasets from computer vision. However, the majority of works related to wildlife focuses only on small datasets (typically subsets of UAV campaigns), which might be detrimental when presented with the sheer scale of real study areas for large mammal census. Methods may yield thousands of false alarms in such cases. In this paper, we study how to scale CNNs to large wildlife census tasks and present a number of recommendations to train a CNN on a large UAV dataset. We further introduce novel evaluation protocols that are tailored to censuses and model suitability for subsequent human verification of detections. Using our recommendations, we are able to train a CNN reducing the number of false positives by an order of magnitude compared to previous state-of-the-art. Setting the requirements at 90% recall, our CNN allows to reduce the amount of data required for manual verification by three times, thus making it possible for rangers to screen all the data acquired efficiently and to detect almost all animals in the reserve automatically. •We address animal detection in realistic UAV imagery (large dataset, vast areas).•We show how to train Convolutional Neural Networks on such a dataset.•We introduce new evaluation protocols suitable for animal censuses.•Based on those, our model yields a precision improvement by orders of magnitude.