Using AI for Conservation in Botswana’s Okavango Delta

This year the team conducted aerial surveys in Botswana, using mounted cameras on the aircraft to take aerial images for analysis. The aerial survey was coordinated not only to provide a baseline index for biodiversity monitoring in the ecosystem but also to contribute to ongoing artificial intelligence (AI) based data processing, providing more images for a range of animals across different habitats. The images contribute to the AI machine learning of, for example, a vulture nest or an elephant. Additionally, the survey provided an incredible platform to train our team to conduct and process these surveys.

This quarter the team, working with a survey lab in Tanzania, spent time tagging the images for machine learning to be able to identify and sort through the aerial survey images and identify animals going forward. During the next quarter, the team will be duplicating the aerial surveys in Botswana to capture dry season data. Next year we hope to be able to take the setup over to our teams in Zimbabwe and expand our use of AI for conservation, as well as set up our own lab to be able to process the images and results of these surveys.

“We were thrilled to be using AI for conservation and capture image of numerous sleeping elephants, not only because it is an unusual sight but also because it meant that these animals were unperturbed by the plane’s presence. Other sightings such as sitatunga, sleeping lions and wild dogs entertained us, while the incredible clarity of images opened up another realm of opportunity with the identification of bird species, something we were not expecting to be able to do with these images Currently, we are collecting additional images of perched vultures, vulture nests and ground hornbills to be able to train the AI software to identify these ‘objects’ in the aerial survey imagery.” – Robyn Hartley, Monitoring & Research