Yu Zhang, Stephen Wistar, Jia Li, Michael A. Steinberg and James Z. Wang,
``Severe Thunderstorm Detection by Visual Learning Using Satellite Images,'' IEEE Transactions on Geoscience and Remote Sensing, vol. 55, no. 2, pp. 1039-1052, 2017. [Download]
You may select a date and a movie will start below shortly after your selection. It can take some time to load the movie. The movie is explained under "Notes" and also in the manuscript.
Select the date:
Storm detection result:
GOES-12 channel 4 image is visualized. The algorithm is running on both channel 3 and 4.
Interval between two frames in the annimation is 30 minutes.
Tinted regions are extracted votex regions (storm candiates).
Red regions are classified as storm regions, green ones are rejected.
If any state is affected by severe thunderstorms in the following hours, it is highlighted in the image.
The highlighting color is in a spectrum from red to yellow. As the color turns to yellow, the correpsonding state is affected by more future storms as far as 6 hours away from the observation.