SPACENET CHALLENGE 7

Multi-Temporal Urban Development Challenge

REGISTER

SpaceNet LLC is a nonprofit organization dedicated to accelerating open source, artificial intelligence applied research for geospatial applications, specifically foundational mapping. SpaceNet is run in collaboration with CosmiQ Works, Maxar Technologies, Amazon Web Services (AWS), Capella Space, Topcoder, the Institute of Electrical and Electronics Engineers (IEEE) Geoscience and Remote Sensing Society (GRSS), the National Geospatial-Intelligence Agency (NGA), and Planet.

THE CHALLENGE

Challenge Overview
Satellite imagery analytics have numerous human development and disaster response applications, particularly when time series methods are involved. The SpaceNet 7 Multi-Temporal Urban Development Challenge aims to improve these methods, while simultaneously advancing the state of the art in SpaceNet’s core mission of foundational mapping. SpaceNet 7 will be featured as a competition at the 2020 NeurIPS conference in December 2020, and seeks to identify and track buildings in a time series of satellite imagery collected over rapidly urbanizing areas. Beyond its relevance for disaster response, disease preparedness, and environmental monitoring, this task poses interesting technical challenges for the computer vision community.

The competition centers around a new open source dataset of Planet satellite imagery mosaics, which will include 24 images (one per month) covering ~100 unique geographies. Each geography features significant urban change over the two-year timespan. The dataset comprises of >40,000 square kilometers of imagery and exhaustive polygon labels of building footprints in the imagery, totaling over 10 million individual building footprint annotations. This Challenge seeks to build upon the advances from SpaceNet Challenges 1, 2, 4, and 6 by challenging participants to automatically extract building footprints and track unique changes at the individual footprint level with computer vision and artificial intelligence (AI) algorithms on this unique time series dataset. This dataset can be found at the SpaceNet 7 challenge overview page (link) or its AWS S3 bucket (link).

Challenge participants will be asked to track localize and individual building construction over time, thereby directly assessing urbanization. What this means is that each building will be assigned a unique identifier (i.e. an address), which must be consistent over time. The ability to uniquely identify buildings in dynamic locales and observation environments (e.g. clouds, haze, seasonality) directly from overhead imagery will be of great utility in one of SpaceNet’s key goals: advancing foundational mapping. Furthermore, timely, high-fidelity foundational maps are critical to a great many domains including disaster response planning and population estimation.

Evaluation Metric
In order to quantify the ability of machine learning algorithms to correctly predict when and where the building footprint change takes place, this challenge will use a new evaluation metric: the SpaceNet Change and Object Tracking (SCOT) metric (link). The SCOT metric combines two terms: (1) a tracking term; and (2) a change detection term. The tracking term evaluates how often the proposal correctly tracks the same buildings from month to month with consistent ID numbers thus measuring the model’s ability to characterize what stays the same as time goes by. The change detection term evaluates how often the proposal correctly picks up on the construction of new buildings measuring the model’s ability to characterize what changes as time goes by.

Algorithmic Baseline
The algorithmic baseline for this Challenge will be made available in August. An explanation of the baseline will be made available on CosmiQ Works’ blog, The DownLinQ (Link).

WHY FOUNDATIONAL MAPPING MATTERS

The commercialization of the geospatial industry has led to an explosive amount of data being collected to characterize our changing planet. One area for innovation is the application of computer vision and deep learning to automatically extract information from satellite imagery at scale. Today, map features such as roads, building footprints, and points of interest are primarily created through manual techniques. SpaceNet believes that advancing automated feature extraction techniques will serve important downstream uses of map data, including humanitarian and disaster response. Furthermore, solving foundational mapping challenges are an important steppingstone to unleashing the power of advanced computer vision algorithms applied to a variety of remote sensing applications.

IN TOTAL CHALLENGE PRIZES

$50,000

SPACENET CHALLENGE 7 TIMELINE

* Timeline subject to slight changes throughout course of challenge

REGISTER FOR THE CHALLENGE TODAY!

REGISTER
NEW TO TOPCODER?

Join the Topcoder Community so you can participate in this amazing challenge and many more like it!

CHALLENGE BEGINS

9/08/2020

MID-POINT CHECK-IN

10/02/2020

CHALLENGE ENDS

10/30/2020

WINNERS ANNOUNCED

12/01/2020