Use of Power Spectral Density Image to Support Early Forest Fire Direct Detection

In this work, we present a method for early forest fire detection from a satellite image using the power spectral density of a Gaussian white noise. To obtain the latter, we have considered each satellite image matrix line as a realization of a non stationary random process in the Thermal Infra-Red (TIR) spectral band, and then divided each line into very small intervals in which any random process can be, obviously, stationary. In addition the pixels of the satellite image are statically independent. Thus, any small interval of each line behaves as a white noise. As a consequence, we have calculated the power spectral densities (PSD) of these intervals and presented them in image form which has been magnified to fix a magnifying factor corresponding to the least hot pixel to appear for early visual fire detection. Key-Words: satellite image; power spectral densities (PSD) image; Thermal Infra-Red (TIR) spectral band; Gaussian model; stationary process; segmentation.