I have watched two documentaries lately about how vast quantities of satellite imagery can quickly be searched for candidate objects for further investigation. On Horizon http://www.bbc.co.uk/iplayer/episode/b01dlglq/Horizon_20112012_Out_of_Control/, Prof Paul Sajda of Columbia University was shown scanning images at a phenomenal rate, with an EEG widget on his head that recognised the unconsious "aha" moment's (as they called it) electonic signature - which they had detrmined from an initial image showing an example of what they were wanting to match. The aplication used for the programme was spotting military bases in Afghanistan - an example of how the best technology and minds are still used for warfare. The commentator opined that future applications of recent scientific discoveries could be used to improve advertising (deep sigh), but more optimistically the massive increase in throughput would help analysis of medical imagery.
On "In Orbit" http://www.bbc.co.uk/iplayer/episode/b01f6qpq/In_Orbit_How_Satellites_Rule_Our_World/ the object being sought was the grave/tomb of Genghis Khan in Mongolia. Not a particularly human need oriented subject, though they didn't mention why they were looking for it, but at least peacable. Doctor Albert Yu-Min Lin of the University of California was approaching this through crowd sourcing - cutting up the images and inviting people to tag the images (sisplayed on the internet) where they see likely structures depicted.
I don't know if the two proffessors know of each other's work, but it strikes me that combining the EEG method with the crowd sourcing method would be extremely powerful.
Incidentally the In Orbit episode on satellites is vey interesing and informative.
No comments:
Post a Comment