Dstl to crowdsource improved satellite imaging solutions
The competition, with a prize fund of $100,000, is seeking to identify novel methods to evaluate large, complex data sets.
Dstl has challenged users to develop an algorithm or software that will automatically detect and identify objects in satellite images. The challenge is to correctly identify and categorise up to 50 features and objects, such as cars, trees and buildings in a variety of environments. The project has the potential to identify solutions which will greatly improve satellite intelligence gathering and analysis capability.
Minister for Defence Procurement Harriett Baldwin said: "The UK has long enjoyed a reputation as a world leader in innovation and it is truly ground breaking projects like the Laser Directed Energy Weapon which will keep this country ahead of the curve.
"The Defence Innovation Initiative and £800 Defence Innocation Fund aim to encourage imagination, ingenuity and entrepreneurship, in pursuit of maintaining a military advantage in the future.
"With a rising Defence budget and a £178 billion equipment plan, our commitment to innovation will deliver a safer and more prosperous Britain."
Entries will be submitted online and will be scored immediately. The results will be summarised on a live leader board. The best three entries will receive a share of the $100,000 prize fund. The first prize is $50,000, the second $30,000 and third $20,000. The competition will run until 28 February and the winner will be announced on 8 March 2017.
Commenting on the competition, Dstl’s Philip Gibson said: "This is a new and innovative way to reach out to a huge community of data scientists who would not ordinarily work with Dstl. We are really excited to be working with Kaggle and their community of ‘Kagglers’ with their well-deserved reputation for exciting, innovative and though provoking solutions."
Kaggle has successfully run over 200 data science competitions. With over 536,000 registered users across 194 countries it is the largest and most diverse data science community in the world.
