Detection of thermal anomalies (fires) from data of AVHRR radiometer and meteorological reports influenced by the atmosphere

摘要

Abstract: A two-stage procedure of detection of thermal, fire-type anomalies on the territory of the region form satellite date of AVHRR radiometer has been considered. At the first stage, the field of thermodynamic surface temperature is retrieved using non-linear, non-parametric regression equation, which is adapted to specific observation conditions using coincident meteoroligical data from the existing meteorological stations as well as data of AVHRR radiometer on NOAA satellite. At the second stage, on the basis of retrieved temperature field, an adaptive Bayes rule of detection of thermal anomalies is constructed, which is based on the principle of identification of the components of mixed distribution and approximation of conditional probability density functions by Johnson curves. An example of identification of thermal anomalies from satellite videodata for the territory of Tomsk region is presented.!6