Internship: deep learning on satellite imagery for the Red Cross
You're passionate about startups and innovation, you're good at programming and you like to work on a project between academic research, startup business and NGO organizations. Yes? Yes? Yes? Continue reading!
Readaar was founded in 2015 and extracts all kind of information from aerial photographs, LiDAR and other sources. To do this we combine remote sensing with machine learning. Our customer base is very diverse: from grid operator to insurance companies. The knowledge we extract from our data has a strong environmental impact: as representative examples, we map solar panels to support the sustainable energy revolution and we help municipalities in banning asbestos. We combine the dynamics of a start-up with the professionalism of our established customers!
The 510 team converts (big) data into understanding, and puts it in the hands of humanitarian relief workers, decision makers and people affected, so that they can better prepare for and cope with disasters and crises. The 510 team is an internationally operating initiative of the Netherlands Red Cross.
The internship assignment is part of a voluntary cooperation between Readaar and 510. Each year, disasters around the world kill nearly 100,000 and affect or displace 200 million people. Many of the places where these disasters occur are literally 'missing' from any map and first responders lack the information to make valuable decisions regarding relief efforts. Missing Maps is an open, collaborative project with a growing number of volunteers to help mapping areas where humanitarian organizations are trying to meet the needs of vulnerable people.
Readaar has developed a deep learning algorithm to automatically map buildings from satellite imagery. Such an algorithm would be really useful to help the volunteers to find buildings easier. Problem is the algorithm is developed for the Dutch situation. In other countries different types of imagery are available and other types of buildings are present. You can imagine that a Dutch house looks completely different than a mud brick house in Africa: therefore the current algorithm won’t work without adaptation.
The information requirements from the 510 team are divers and covers areas (e.g. flood extent), lines (e.g. roads and waterways), points (e.g. trees divided by type of trees) and attribute information (e.g. roof type and construction materials). This information is now collected through visual interpretation and GIS analysis through separate projects and through (public) missing maps sessions.
During this assignment you will start with the existing algorithm and adjust this to make it more general applicable. The 510 team will supply validation data from previous mapping projects, as well as testing imagery. The first goal is to reduce the amount of manual work needed and to speed up the process during manual mapping. This will be mainly focused in this pilot on disaster preparedness areas. The ultimate goal is to be able to use this for disaster response when every hour counts.
About you and the project:
Your focus within this project is on the automatic and accurate identification of buildings. The main challenges to be solved will be in the variety of buildings to be detected and different types of satellite imagery. To make this a success:
Can't wait to be part of a small, fast growing innovative startup. Convince us and hopefully we can welcome you as our new team member.
For more information call Sven Briels +31 (0)6 289 14 981, or apply via email@example.com
We just sent you an email. Please click the link in the email to confirm your subscription!
OKSubscriptions powered by Strikingly