Detection and classification of running vehicles based on acoustic signatures

摘要: A method and apparatus for identifying running vehicles in an area to be monitored using acoustic signature recognition. The includes input sensor capturing waveform produced by a vehicle source, processing system. is digitized divided into frames. Each frame filtered plurality of gammatone signals. At least one spectral feature vector computed each frame. vectors are integrated across frames create spectro-temporal representation the waveform. In training mode, values from used as inputs Nonlinear Hebbian learning function extract signatures synaptic weights. active weights patterns supervised associative network identify whether present monitored. response being present, class identified. Results may provided central computer.