Satellite technology and deep learning are being used in the conservation of Africa’s iconic species, including the giraffe.
In Zimbabwe’s Hwange National Park a team of French scientists has developed a deep learning computer system to recognise the coat patterns – brown blotches on a tan background – and thus distinguish between individual animals. Populations in this remote part of north-western Zimbabwe are on the decline, and the scientists are using the technology to find out why. “To our knowledge, this is the first attempt in using deep learning techniques for this task,” lead author of the new study, Vincent Miele from the University of Lyon’s Laboratory of Biometry and Evolutionary Biology told RFI.
Between 2014-2018 his team photographed around 400 giraffes in Hwange, the park where Cecil the Lion once lived. Out of a set of nearly 4,000 pictures a training dataset was created by cropping the images to display the animals’ flanks. These were fed into a computer system, known as a convolutional neural network (CNN) that analyses visual imagery.
This deep learning system, mostly used in facial recognition of humans, was trained by the scientists to re-identify individual giraffes in Hwange with 90% accuracy. “The advantage of deep learning is that once the computer has been trained, it is very fast and can tackle dozens of images in a few seconds,” explained Miele. “Deep learning’s algorithms are known to outperform any other algorithms in terms of predicting performance.”
Giraffes have undergone worrying population declines without the headline-grabbing attention given to other megafauna, like lions, elephants and rhinos. The International Union for Conservation of Nature recognises nine subspecies of giraffe. Most are listed as near threatened, vulnerable, endangered or critically endangered. The Giraffe Conservation Foundation an independent group working in 16 African countries, estimates the animals’ continent-wide population has declined by nearly 30% since the 1980s, from more than 155,000 to around 117,000 now.
Although the CNN system used in Zimbabwe was designed to recognise and monitor giraffes, it uses freely-available software and researchers can tweak it to apply to a range of other mammals. In Hwange, this might include zebra and kudu, a large antelope that has majestically twirled horns and, critically, unique stripes on its coat. (www.rfi.fr 12/05/2021)
Rhino Footprints: Interactive software is also being used to monitor endangered black rhinos in Namibia. The software, called the Footprint Identification Technique (FIT), uses advanced algorithms to analyse more than 100 measurements of a rhino’s footprint. Because each rhino’s footprint is as distinctive as a human fingerprint, the analysed images can be archived electronically in a global database of previously collected footprint images for matching. A team from Duke University’s Nicholas School of the Environment is working with Namibia’s Ministry of Environment, Forestry and Tourism to train wildlife conservationists, land managers, local guides and anti-poaching agents how to use FIT.
Namibia is home to an estimated 2,000 black rhinos, or about 90% of the species’ total population worldwide. Stepped up government policing in recent years has significantly slowed the rate of loss due to poaching, but between 30 and 50 of the animals are still slain each year for their horns, which can sell for more than $60,000 a kilogram on the Asian black market, where they are used in traditional medicine or displayed as a symbol of wealth and success. (www.sciencedaily.com 09/09/2020)
Counting Elephants: Meanwhile a team from Oxford University’s Wildlife Conservation Research Unit (WildCRU) the University of Bath and the University of Twente working in South Africa is counting elephants from space using satellite imagery and automating detection via deep learning, having built up a dataset of some 1000 animals. The population of African elephants has plummeted over the last century due to poaching, retaliatory killing from crop raiding and habitat fragmentation.
Existing methods are prone to error. Inaccurate counts lead to misallocation of scarce conservation resources and misunderstanding population trends. Satellites can collect upward of 5000 km² imagery in one pass captured in a matter of minutes, eliminating the risk of double counting. Repeat surveys are also possible at short intervals.
Satellite monitoring is an unobtrusive technique requiring no ground presence thus eliminating the risk of disturbing species, or of concern for human safety during data collection. Previously inaccessible areas are rendered accessible, and cross-border areas – often crucial to conservation planning – can be surveyed without the time-consuming requirement of terrestrial permits. The scientists first developed the techniques at South Africa’s Addo Elephant National Park. (www.ox.ac.uk 18/12/2020)
Habitat loss is one of the main drivers of species depletion. In Tanzania, The Nature Conservancy (TNC), is monitoring forest loss in the Greater Mahale ecosystem and Lake Tanganyika water basin with satellite imagery, comparing it with historical photos. It found that forest cover in the Lake Tanganyika basin has declined by about 26% over the past three decades, and forest loss in the Greater Mahale ecosystem by about 10% in the same period. East Africa is estimated to have lost nearly 15m acres of forest between 2000 and 2012 for reasons including timber harvesting and agriculture.
Between 2015 and 2018, TNC used satellite imagery to map the entire country of Zambia, in partnership with the government, to identify where animals such as giraffes and hippos congregate and create a species distribution model. Satellite imagery was overlaid with aerial surveys from planes to identify the animals. (www.devex.com 05/06/2020)
Further Reading:
African Wildlife Foundation: https://www.awf.org
WWF Africa: https://africa.panda.org
The Nature Conservancy Africa: https://www.nature.org/africa
Africa Research Online 