Early fire detection using non-linear multitemporal prediction of thermal imagery

Multitemporal anomaly detection method called Dynamic Detection Model (DDM) shows superior performance when detecting hot spots in MODIS images.

Temporal dimension of the MODIS image sequences is more informative for wildfire detection than spatial context.

More Details:

Koltunov, A. and  S.L. Ustin (2007) “Early fire detection using non-linear multitemporal prediction of thermal imagery”, Remote Sensing of Environment, 110, 18–28. (View PDF)