Out of my window (Aurora Globe)

Develop virtual reality viewers using International Space Station video feeds, Aurorasaurus or other camera data that features views of the northern lights, or gives users other virtual experiences of viewing the Earth from the International Space Station.

How we worked on the task: The team members who got together on this task didn't start with a well formed goal, so much of the team's time initially was spent learning and thinking about the opportunities with aurora pictures taken outside the astronaut's window. We wanted to look at better ways of visualising/using the actual images. We looked at the work done with aurora 'weather' prediction (http://www.aurora-service.eu/aurora-forecast/), aurora localisation (with http://aurorasaurus.org/) and looked through github material referencing the ISS for inspiration. We thought about the problem involved with working out where images were taken and looked at (http://www.cosmos.esa.int/web/arrrgh) and discovered just how hard this was. We were impressed by the complex work done to correct and project the imagery (lens distortion, time correction using earth horizon and star field photogrammetry...) and we thought we couldn't add to that work but we could contribute by trying to improve the way in which the images were visualised. In particular we decided it would be interesting to try to re-project the localised/mathematically reprojected aurora pictures onto a virtual globe so that it could be more readily interacted with.

How we built this: we built a simplified globe with an earth projection texture and made use of javascript libraries to allow the rendering to happen in browser. We used an available earth texture, but had wanted to use the NASA blue marble images with correct months as textures and to project a day-night shadow (but we ran out of time and had prioritised getting something to show the idea). We had ideas for extensions such as having different layers onto which the different light wavelengths of aurorae would be projected (we discovered in reading about them that different colours are typically present at different atmospheric heights as the cascade of excitation descends). We struggled to get the ESA software running and this became a big issue when we realised there were issues with the shape of the images and we wanted to get them correctly placed into our aurora projection layer texture. Having wrestled with installation on three different versions of linux and battled with erratic network availability, we were disappointed to have to opt to manually place an aurora field into the aurora texture layer to show the concept by the deadline - we make no claim about the accuracy of scale and position for the image, but tried to do align it using lat/long info in an image editor.

Further ideas: we did have ideas such as using the ISS location information to calculate potential periods when the ISS would pass through a dark zone near the poles (where visual contrast would be best) and software might activate recording of the video capture from the ISS's HD streaming camera, perhaps also integrating aurora weather prediction from existing sites using ACE satellite data, or notifying astronauts of the potential view. We thought simple image processing for colour could also have identified periods for suitable image capture.

These are resources consulted/used in addition to those mentioned above, but not all of this is evident in output!

GoogleDocs - to collate information which we researched online.

The resources provided by https://github.com/SpaceApps2016/Resources

Github - we searched through packages including "ISS"

Other NASA things:

NASA live ISS Ustream - though this proved unavailable through long periods during the hackathon!

NASA GIBS - for inspiration of data visualisation

European Space Agency, in particular the 'Automatic Georeferencing of Astronaut Auroral Photography' http://www.cosmos.esa.int/web/arrrgh/software - but their softwware proved a problem to install - we would suggest a docker image (in the style of the hub.docker.com/r/pkgw/jupyter-py2-astrostack)