Thermal remote sensing is a valuable tool for monitoring active volcanoes. One can use it to detect smaller thermal anomalies, such as hot springs or fumaroles. It is successful also by detecting larger anomalies, such as lava flows or lava lakes. All these anomalies usually cover only a small area comparing to the ground sample distance of the sensors that are appropriate for their monitoring. This makes detecting and characterizing the thermal anomaly a difficult task that is solved by using more data from several bands covering the appropriate spectra. The most robust parameter to characterize a thermal anomaly is radiative heat flux. We need to analyse its time series in order to understand the volcano activity. Satellite data are influenced by the atmosphere, terrain, etc. Therefore, the time series always contains some noise. For a case study we applied Kalman filter to minimise the noise in the case study of Etna eruption in autumn 2002.