The recognition method of thunder based on a noise-estimation algorithm for highly non-stationary environments

A recognition method of the thunder signal based on the end-point detection method with a noise-estimation algorithm for highly non-stationary environments is proposed. To determine the parameter of the algorithm and verify the effectiveness and accuracy of the method we analyze the acoustic signals obtained from two microphone arrays installed at the top platform on a seven-floor building in Wuhan. In this paper, we have chosen two acoustic signals to test the endpoint detection method. It is revealed from the recognition result that the method can find out the accurate beginning and ending time of the thunder signal from the acoustic signals.