To aid in the protection and conservation of the environment and its wildlife, the FruitPunch AI community aims to develop an autonomous UAV based solution to revolutionize the way monitoring and inspection of wildlife reserves is conducted.

Buster Franken - AI for Wildlife initiator Tweet

The context

Most of you will know about the poaching crisis in Africa. But you may be asking yourselves: why is it especially important now? With the middle class of many Asian countries growing, the demand for the luxury erection-enhancing, cancer-curing powder of rhino horn, ivory and pangolin scales is growing. Of course, these don”t actually have any of the previously mentioned properties. After all, rhino horn is made from keratin, the same stuff your nails and hair are made from. But that doesn’t stop the international black market from craving the product, driving up the price of the endangered animals’ body parts higher than the price of gold, and motivating the poorest (or most corrupt) people to kill these beautiful animals by the ten-thousands.

The real solution here is to stop the demand by educating the masses that eating or keeping these body parts doesn’t do anything but make you an asshat. But unless we can stop this disease within the next 10 years, which is highly unlikely, many species will die out. Thus, we also need to stop the bleeding. This means applying our best selves and our best tools to stop poaching in the wildlife reserves. And that is what we set out to do, using humanity’s most powerful tool, artificial intelligence.

The FruitPunch AI Community is working together with SPOTS – Strategic Protection of Threatened Species, who represent the volunteers and rangers of Pilanesberg National Reserve, to tackle the poaching crisis in their home territory.

They are experienced wildlife conservationists that have been using drones with thermal camera’s to search for poachers in the night when they are most active.

They are facing the following challenges:

🍉   Flying time and range are limited with RC planes
🥝   Specialized and skilled pilots for landing RC planes are required 
Watching the black and white video stream for poachers through the night is prone to human error

🍒   Most drones can’t navigate where GPS is inaccurate or unavailable 
🍎   The application environment is characterized by strong dynamic external disturbances, such as air force
There is limited flying time due to battery capacity

In the first AI for Wildlife challenge, a team from the community developed a detectronV2 model to detect poachers on the thermal video feed of the drone, & they’ve helped SPOTS to collect higher quality data. Now, we’re onto the second challenge, creating a ML model that can detect poachers and run on-edge on the drone!

Watch the final presentation of the first challenge here!

The team of challenge #1:

Thank you Matthew Lewis, Mert G., Bram CalsAbhishek VenkatachalamAnwar AmezougDaphne Smitsfaiz ikramulla ⚙️Iris den HartogIshwarya ChandramouliJordi FrankKamalen ReddySmahane RziakAdrian – Daniel AzoiteiCecilia LiuKhaoula El AhmadiLuke IgnjatovićSako Arts, Nur Y. Nick van Santen!

The Second Challenge

The second AI for Wildlife challenge concerns: Creating a ML model for on-edge detection of poachers on the thermal video feed of a wildlife protection drone!

You will be working on one or several of the following topics in this challenge:

  • Object recognition on thermal video footage from a wildlife protection drone
  • Scaling down the model to implement it on-edge (on-edge == on the hardware on the drone)
  • Hardware design to run the ML model on the drone without guzzling up too much energy, taking up too much space or weighing too much 

To learn more, rewatch the info session given by us & the conservationists HERE!

Who we’re looking for

You can apply at any level of experience above basic (theoretical) coding & data science skills. If you are just starting out, join the platform & turn on email notifications because we will be releasing an AI Bootcamp soon! In the selection of the team we look for a combination of rookies looking to learn, hardened professionals and life-long learners switching it up. We are looking for:

  • AI engineers / data scientists > experience with object recognition is a pro!
  • Electrical / mechatronics / mechanical engineers with a nack for AI implementation

You can join as a contributor (12 hours per week commitment for 2 months), coach (2-4 hours per week, only for experienced ML professionals) and teacher (give one relevant ML / domain masterclass).

During the challenge we will arrange for masterclasses on relevant topics like the use of the cloud computation resources Microsoft, IBM & the HPC Lab have made available for us. You will also build upon the knowledge and experience of the first AI for Wildlife’s challenge team!

To learn more, rewatch the info session given by us & the conservationists HERE!

Challenge format

The challenge will run from the 15th of September untill the 17th of November, & you will collaborate with a diverse team of over 30 international data specialists and domain experts in subteams, all tackling this problem from different angles.

Some important dates:

  • 18th of August – Info session
  • 15th of September – Challenge Kick-off
  • 6th of October – Midterm presentations
  • 17th of November – Final presentations

Apply to join the Challenge

The deadline for application is 13 September 23:59 CEST

Fill in the Google Form behind this button to become one of the 50 engineers collaborating with SPOTS to develop ML models that can detect poachers in the wildlife reserves of Africa!!

Quick facts

🥥  More than a thousand rhinos are slaughtered per year for their horns.

🍊  More than 100,000 african elephants were killed between 2014 and 2017 for ivory. And there are less than 4000 tigers in the world.

🍉  In Africa, almost 600 rangers charged with protecting wildlife were gunned down by poachers between 2009 and 2016 while working.

🥝  In the Democratic Republic of the Congo’s Virunga National Park, at least 170 rangers have been killed during the past two decades.

Source: National Geographic

Source: Wildlife justice


Strategic Protection of Threatened Species
Learn more