I wish the article was more clear what advances the Berkeley folks have come up with. They claim a theoretical 3 meter resolution compared to the 'next-gen' sensor package (VIIRS [1]) at 750 meters.
I had the same question. From "Table 1" of the linked article, they have 72m X 72m pixels, and they are trying to detect a 3m X 3m fire (burning at 1100K) within that pixel.
To bring up a related matter, it seems like they are not proposing to do the detection on-board (see section 2.2.2, where they talk about a database of past images). If they are not, they have to transfer all the data to ground. The data rate can be very demanding (and the high orbit makes it more so). Accurate pointing will be another issue (high orbit, small pixels). They have a longer way to go than the press release indicates.
Notwithstanding all this griping, it's an interesting concept.
I wish the article was more clear what advances the Berkeley folks have come up with. They claim a theoretical 3 meter resolution compared to the 'next-gen' sensor package (VIIRS [1]) at 750 meters.
[1] http://npp.gsfc.nasa.gov/viirs.html http://collaboration.cmc.ec.gc.ca/science/rpn/SEM/dossiers/2...