Header Slideshow Image 1 of 13
Header Slideshow Image 2 of 13
Header Slideshow Image 3 of 13
Header Slideshow Image 4 of 13
Header Slideshow Image 5 of 13
Header Slideshow Image 6 of 13
Header Slideshow Image 7 of 13
Header Slideshow Image 8 of 13
Header Slideshow Image 9 of 13
Header Slideshow Image 10 of 13
Header Slideshow Image 11 of 13
Header Slideshow Image 12 of 13
Header Slideshow Image 13 of 13


Due to their large ranges and mobility surveys of whales present a significant challenge. Worldwide, there is tremendous interest in learning how whale populations are recovering following the end of intense commercial whaling which lasted until the second half of the 20th century, considerably decimating many populations.

Observations and counts make use of the fact that large numbers of these animals reliably concentrate in particular areas at key times (as for example the mating and birthing season). Aerial surveys are an effective and common way to locate whale populations and track large-scale movements but can be costly and are restricted by aircraft range.

Humpback whales bubble-net feeding in Canada

In collaboration with ocean ecologists of the Stony Brook University in New York and HiDef Aerial Surveying Ltd from the UK, BioConsult SH has created a semi-automated process for whale detection from very high-resolution satellite images using deep learning artificial intelligence. The method was successfully tested using images from areas known to host whales off the Hawaiian and Argentinian coast.

The algorithm was trained using down-scaled HiDef digital aerial images and satellite images to classify whether a tile image was likely to contain a whale. The best model correctly classified 100 % of tiles with whales, and 95 % of tiles containing only water.

HiDef image of a Minke whale

While the relatively poor resolution of commercially available satellite images continues to make whale identification a challenging problem, our approach provides the means to efficiently eliminate areas without whales and, in doing so, greatly accelerates whale surveys.

Further information on this project can also be found on the website of the European Space Agency (ESA), which has funded the project.

The method has been described in a PLoS ONE article which can be downloaded here:

Borowicz, A., Le, H., Humphries, G., Nehls, G., Höschle, C., Kosarev, V. & Lynch, H.J.
Aerial-trained deep learning networks for surveying cetaceans from satellite imagery.
PLoS ONE 14(10): e0212532
Download documentExternal Link

The project is funded by the European Space Agency (ESA).


Caroline Höschle