A new algorithm for forest fire detection method with statistical analysis using NOAA AVHRR images

Abstract

Various researchers have carried out forest fire analysis using NOAA satellite images. There are several methods of doing this, and most can detect a fire. However, many false fires were also detected, and, in some cases, actual fires were missed. We analysed four satellite‐based fire detection methods using data from AVHRR of NOAA‐16 over a period of three to six months for the Sakhalin region and the Japan region. Considering the fundamental differences, problems, and effectiveness of these methods, we have constructed an improved fire detection method with statistical analysis. Our method has reduced false fire detection significantly, as well as detected actual fire with accuracy.