Time-frequency features extraction from acoustic response waveform of marine sediment and its intelligent classification

Applied modern signal analysis method, four characteristic parameters have been extracted from acoustic response waveform of marine sediment: sound velocity C_p, index of amplitude ka, waveform relation dimension fraction index D and frequency moment of acoustic spectrum K_f. Build neural network model, with these four characteristic parameters as input vector and structure type of marine sediment as output vector. Examples prove that the neural network model built in this way has strong classification and capability.