An Enhanced Algorithm for Forest Fire Detection Based on MODIS Data

The theory and methods of MODIS (Moderate Resolution Imaging Spectroradiometer) data in fire detection are introduced. Experience with high quality data from MODIS has suggested several improvements to the original active fire detection. An enhanced algorithm using variance between-class and smoke plume mask is described. The brightness temperature threshold of potential fire pixels was adjusted to be 305K. Based on the variance between-class of thermal infrared spectral channels, hot fire spots and cool fire spots can be separated from the background respectively. Smolder spots of low temperature were distinguished with algorithm of smoke plume mask. It has been used in forest fire detection happened at Fujian province and Blagoveshchensk of Russian. It satisfy with different environments and suit to identify high-temperature fire points and low-temperature smolder points accurately.