Close

Key blogs, articles, and interviews to help you compete and succeed in the SpaceNet Challenge on Topcoder

 The SpaceNet Challenge

Asset Library


 

DigitalGlobe, CosmiQ Works, and NVIDIA are challenging the Topcoder Community to develop automated methods for extracting building footprints from high-resolution satellite imagery. 


Applying computer vision techniques to automate the extraction of buildings from satellite imagery will help create more accurate maps, more rapidly.

The SpaceNet Challenge Will Help Create Next-Gen Computer Geospatial Vision Algorithms Utilizing Real Satellite Imagery and Data

The teams from DigitalGlobe,  CosmiQ Works, NVIDIA, and Topcoder have created this useful asset to make it easy for you to better understand this topic and to prepare to compete at a high level in The SpaceNet Challenge. If you are new to the area of geospatial computer vision algorithms, we suggest you explore the following content and plan to attend a Google OnAir Hangout featuring experts from the SpaceNet team that we will be hosting prior to the launch of the challenge. To pre-register, go to this page and fill out the simple form. Enjoy the content and good luck! 

Getting Started with SpaceNet Data

Winning Implementations and New Imagery Release

Building Extraction with YOLT2 and SpaceNet Data 

How will The SpaceNet Challenge be scored?

Object Detection on SpaceNet

SpaceNet Challenge Utilities GitHub Assets Repo 

The SpaceNet Challenge Visualizer

Here is the SpaceNetâ„¢ Challenge Visualizer - Explore this important tool today!

DigitalGlobe, CosmiQ Works, NVIDIA, and Amazon Web Services Team up to Launch SpaceNet Open Data Initiative

TechCrunch: SpaceNet 

Satellite Imagery Repository Launched 

SpaceNet Experts OnAir Google Hangout Q&A

Exploring the SpaceNet Dataset Using DIGITS

The Spacenet objective is to build a large database of satellite imagery and to host prize-based competitions to foster new analysis techniques. In support of this objective, Spacenet has an Amazon Web Services S3 repository to make the data accessible. SpaceNet imagery and associated labels are licensed via the Creative Commons. Spacenet plans to release more data to the repository with a planned pipeline of over 150GB.

Want to compete in this unique algorithmic challenge on Topcoder?