AI for Earth Challenge
Using European Space Agency Earth observation data to show humanity’s impact on planet Earth.
Map humanity's impact on earth using ESA’s satellite imagery
It seems like nowadays we can’t go online without seeing bad news about our natural environment. With the fires devouring Greece we couldn’t help but wonder about what we can collectively do to prevent this. Due to human activity and built up fuel, wildfires have increased over 33-fold in places like Greece. We know early warning systems have to be a part of the solution, and ESA’s Earth Observation satellites provide us with a detailed view of our planet. The AI for Earth challenge will focus on creating machine learning models that can detect & prevent negative impact on our natural world.
FruitPunch AI community in collaboration with the European Space Agency - ESA, will use its Observation Image Data to create CNNs and focus on anomaly detection & contrastive learning to detect human impact on Earth and guide conservation efforts. ESA will provide us with multi-spectral data of their Sentinel-2 satellite pair and with valuable knowledge and research on the domain of Earth Observation data in participant only masterclasses.
Our challenge partners
GOAL: Create ML models that can detect humanity’s impact on planet earth & fight climate change
The primary goal is to get an insight into the human impact on Earth, to drive and guide conservation efforts of this planet we call home. The entire team of 50 AI engineers will be split into working groups focusing on:
- Illegal deforestation
- Detecting wildfires
- Air quality detection & prediction
- Coral reef monitoring
- Dying rivers
Our approach will be twofold:
- First, we will work on algorithms that can serve as an early detection system of human impact sites.
- Next, we will use these detection systems to find satellite images that show the most impactful human-caused changes, which will be used to create a video to launch an awareness campaign.
AI for Earth Challenge Info Session
Date: 3. August 2021
Presenter: Nicolas Longépé, Earth Observation Data Scientist at ESA
Who are we looking for?
During this challenge, you will have the opportunity to use Observation Image Data to create CNNs and focus on anomaly detection & contrastive learning to detect human impact on Earth. If you’re somebody with an engineering background, a love for space, and an interest in saving our planet, this challenge is for you!
We are looking for engineers with some experience with programming interested in machine learning. Anyone with the right motivation and ‘proof’ of understanding of the core concepts found in the application form can sign up!
You can join as a contributor (8-12 hours per week commitment for 10 weeks), coach (2-4 hours per week, only for experienced ML professionals) and teacher (give one relevant ML / domain masterclass). The challenge will run throughout September and October 2021, where you will collaborate with a diverse team of over 50 international data specialists and domain experts in subteams for illegal deforestation, detecting wildfires, air quality, coral reef monitoring and dying rivers.
Over the course of the challenge they will receive masterclasses from:
- ESA and Space4Good about working with open-source Earth Observation data
- MLReef about creating & maintaining an ML pipeline
- Adobe about creating an impactful video from the results
Application Deadline: 31 August 2021
Challenge Kick-off: 1 September 2021
Midterm Presentation: 29 September 2021
Final presentations: 29 October 2021