Vol. 9, issue 05, article # 10

Abstract:

The feasibilities are explored for detection of seats of forest fires using spectral infrared radiation measurements under conditions of cumulus clouds from space. This problem is solved by a statistical hypothesis testing method. A decision rule is based on the Neumann–Pearson test. Statistical models of signal (the fire seat emission) and background (the radiation from the underlying surface, clouds, and the atmosphere) have been proposed and validated. Analysis of the results obtained shows the feasibility of reliable detection of forest fires over an area of about 600 m² provided that the cloud amount does not exceed 0.5.