SAVING ELEPHANTS FROM THE SPACE USING AI

- Written by Deva Kumar, AI Alchemy by Orane

Welcome peeps to the first edition of ‘Interesting discoveries in AI’. There are many more awestruck inventions and advancements in the world of artificial intelligence(AI) on the cards. So, stay tuned for the weekly updates.

Human beings are the only creature in the face of the world which has the potential to create, destroy and save lives. And the only creature which does not get satisfied with what it has and always wanting more. It has equally invented and innovated new things and has destroyed the existing things in this whole world.

Now, it has come to a point where destruction has outweighed creation. It is up to humans now, whether they want the world to be like the romances or like the tragedies of William Shakespeare. More than a necessity, humankind needs to be accountable and responsible. Can you guess the largest land animal on Earth? Yes, you guessed it right. It’s iconic and colossal African Elephants. It is time to bid farewell to the species as they are set to become extinct by the year 2040 and have been listed as vulnerable to extinction by the International Union for Conservation of Nature(IUCN) due to poaching and wildlife destruction.

The saddest part is that elephants have evolved to be born tuskless due to high levels of ivory poaching. Accurate monitoring is crucial at this stage in order to protect the last 400,000 African Elephants on planet Earth. Researchers led by the University of Oxford developed a survey method combining high-resolution satellite imagery and artificial intelligence(Machine Learning) to detect and track the elephants in a heterogeneous landscape of dense African forest.

So, how does AI plays its part in the survey method?

Let’s see how it’s done. The researchers developed a Convolution Neural Networks(CNN), a deep learning algorithm to process and automate the detection of elephants in the captured satellite images. A training dataset of nearly 1000 elephants was collected from the Addo Elephant National Park in South Africa to provide information on where the elephants are and how they look from space. Machine Learning(ML) was applied via a CNN to train the model in order to filter out and learn the difference between elephants and heterogeneous landscape(trees and grasslands). Using satellites WorldView-3 and WorldView-4, this survey method is quick, accurate and efficient in comparison to aerial surveys and camera- trap surveys.

With a continuous improvement in the resolution of satellite imagery and advancement in the field of artificial intelligence, the data obtained is going to be a game-changer in tackling wildlife challenges and key issues in wildlife conservation.

Credits to the researchers: Isla Duporge, PhD Researcher, University of Oxford. Dr. Olga Isupova, Postdoctoral Research Associate, University of Oxford. Dr. Steven Reece, Dept. Research Fellow in the Pattern Analysis, University of Oxford.

Interested in learning about artificial intelligence and how the algorithms of machine learning can be used accurately and effectively in real-time monitoring.

Check out the tailored and live artificial intelligence courses offered at https://aialchemytraining.com/. Book a demo class today and experience the learning. If you like our content, kindly share it with your friends and learn Interesting Discoveries in AI together. Thanks a million for your time.

References: Baker, H., 2021. Elephants counted from space using satellites and AI. [online] livescience.com. Available at: [Accessed 12 February 2021].

Imagine, E., 2021. Using Satellite Imagery and Machine Learning to Detect and Monitor Elephants. [online] Sensing Change Blog. Available at: [Accessed 12 February 2021].

The Engineer. 2021. Scientists count elephants from space using AI | The Engineer. [online] Available at: [Accessed 12 February 2021]