摘要: Acoustic imagery generated by the Naval Research Laboratory's ASCS (Acoustic Sediment Classification System) is analyzed using time-frequency methods. The ASCS Imagery is generated using normal-incidence acoustic data. Nonlinearit... 展开 Acoustic imagery generated by the Naval Research Laboratory's ASCS (Acoustic Sediment Classification System) is analyzed using time-frequency methods. The ASCS Imagery is generated using normal-incidence acoustic data. Nonlinearities in the propagating medium produce multiple harmonics in the return. Different components of the bottom and sub-bottom produce different spectral signatures, based on the harmonics that can be used to classify the bottom. Although the exact acoustic properties of the media may not be known, it is possible to segment the spectrally different components. This paper presents an efficient method for performing the time-frequency analysis. The tradeoffs between time and frequency resolution are examined and exploited to produce the best overall resolution. Examples using the ASCS data demonstrate its use for classification of acoustic imagery. Buried pipes, methane gas deposits, and other features can be differentiated using this method. 收起