Computer Vision for the Solar Dynamics Observatory: Status and What's Next?

Martens, Petrus

NASA has selected our large international consortium in 2008 to produce comprehensive system for automated feature recognition for the imagery of the Solar Dynamics Observatory (SDO). We are producing robust and very efficient software modules that can keep up with the SDO data stream of about 1.5 TB per day, and detect, trace, and analyze a large number of phenomena, including flares, sigmoids, filaments, and coronal dimmings. I will demonstrate results from the several modules that have been inaugurated since the end of SDO commissioning last summer.

I will then describe our future plans for this project, in particular those for the trainable automated feature finding

module.

Finally I will present my vision of the development of the "Emerging Informatics Infrastructure " in Solar Physics, as a convergence of the development of Virtual Heliospheric Observatories and Computer Vision applications.

Return to invited speaker list