Dark Skies ISS image challenge



More than 100,000 pictures are lost in the ISS archiveā€”help us to classify them.

At the Loss of the Night Network and the Light pollution group of Universidad Complutense de Madrid, we want to do science with the night-time images taken from the International Space Station. But they are lost in the ISS archive. We need geotagged images to use as a tool against light pollution and energy waste.


We’d like your help with two challenges.

Challenge #1: Automated classification

We need a automated classification of the day and night images based on the time the images were taken. Also a visual classification to distinguish the quality of the images and which are from stars, cities or other strange things.

Challenge #2: Crowdsourcing interfaces for image georeferencing

We need to georeference more than 10,000 images so we need people to do it and a good interface.

In the first stage, people need to make a first guess of the position of the center of the picture. To do that we can use the Google gallery and the Earth at Night 2012. But these images don’t have a very good resolution. So this will be useful mostly for the wide angle pictures. To georeference the high resolution images we can use the Google maps pictures or the Open street maps.

Then, we need a high-quality georeference, so the citizens will identify control points in common in the image and in the map. Here we need to capture the physical and image coordinates of 10 control points.

You can drop by our table at the Hack Day to learn more about our challenge and our project.

Find us on twitter at @pmisson and @skyglowberlin. Our hashtag is #NightISS