Deep learning based whale detection from satellite imagery

Kapoor, Saakshi; Kumar, Mukesh; Kaushal, Manisha

doi:10.1016/j.suscom.2023.100858

Cited by 6 publications

(1 citation statement)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Object detection with deep learning has been successful in identifying whales in VHR satellite imagery (Houegnigan et al., 2022; Kapoor et al., 2023; Khan et al., 2023). Training of CNNs has been previously performed using aerial imagery or a mixture of VHR satellite and aerial imagery (Borowicz et al., 2019; Guirado et al., 2019), due to the lack of labelled VHR satellite imagery, which takes time to acquire and must be manually requested and annotated by users (Cubaynes & Fretwell, 2022; Höschle et al., 2021).…”

Section: Introductionmentioning

confidence: 99%

Gray whale detection in satellite imagery using deep learning

Green

Virdee

Cubaynes

et al. 2023

Remote Sens Ecol Conserv

View full text Add to dashboard Cite

The combination of very high resolution (VHR) satellite remote sensing imagery and deep learning via convolutional neural networks provides opportunities to improve global whale population surveys through increasing efficiency and spatial coverage. Many whale species are recovering from commercial whaling and face multiple anthropogenic threats. Regular, accurate population surveys are therefore of high importance for conservation efforts. In this study, a state‐of‐the‐art object detection model (YOLOv5) was trained to detect gray whales (Eschrichtius robustus) in VHR satellite images, using training data derived from satellite images spanning different sea states in a key breeding habitat, as well as aerial imagery collected by unoccupied aircraft systems. Varying combinations of aerial and satellite imagery were incorporated into the training set. Mean average precision, whale precision, and recall ranged from 0.823 to 0.922, 0.800 to 0.939, and 0.843 to 0.889, respectively, across eight experiments. The results imply that including aerial imagery in the training data did not substantially impact model performance, and therefore, expansion of representative satellite datasets should be prioritized. The accuracy of the results on real‐world data, along with short training times, indicates the potential of using this method to automate whale detection for population surveys.

show abstract