Dark Skies ISS

Image pattern recognition



published results
tasks done
pending tasks

Note: Interface in english for not registred users, we are working on the translations contact here. Descripción en Español Description en français

Right now there are around 1,800,000 images at the Johnson Space Center database (The Gateway of the Astronauts). Around 1,200,000 images were taken aboard the ISS (date 02/20/2014). However, the number of classified images is much smaller, and there is no archive of georeferenced images. There is already a project to classify the daytime images (Image detective), but the techniques used in that project are not useful for the classification of nighttime images. The patterns on Earth are not the same during the day and night, which is why another technique is needed to classify these nighttime images.

Our main objective is to study light pollution that comes from cities. We want to stop the waste of energy and the destruction of the mighty ecosystem.

Your collaboration is really important because algorithms cannot distinguish between stars, cities, and other objects (i.e. moon). Thus, we need your help to assess the light pollution in our world!

For more information, please contact www.citiesatnight.org or Twitter handle: @cities4tnight.

Also you can contribute on our other apps Lost at night (find unlocated images) and NightCitiesISS (Georeference known cities). Also you can track the project's updates at our blog.

The Cities at Night website content was developed entirely to be used for outreach and scientific research. Its content and all derived products created as result of Cities at Night project activities can be used for those purposes as long as the following citation of the project is included:

Creative Commons License. Cities at Night by Alejandro Sánchez de Miguel et al. Atlas of astronaut photos of Earth at night A&G (2014) is distributed under Creative Commons Non commercial attribution – Share 4.0 International licence.

However, if the Cities at Night data are a principal component of a science paper then co-authorship to PI’s should be offered.

Icon of the application courtesy of NASA/ESA JSC


UCM Campus MediaLab LoNNe REECL

Images thanks to the astronauts of: NASA,ROSCOSMOS,ESA,JAXA,CSA-ASC

Promotion with help of: NASA ESA CANADA