A codebook of feature vector for underwater targets

Sound is one of the few forms of energy that will propagate reliably underwater, and for this reason it is used by aquatic animals for navigation and communication. Man also uses sound for the same reasons underwater, and additionally generates noise as a byproduct of offshore activity. The problem of identification of noise sources in the ocean is of great importance because of its varied practical applications. Hidden Markov Models can provide an effective architecture for classification of underwater noise sources. In this paper, a methodology is presented for the Code book generation of feature vectors extracted from under water signals. The generated Codebook can be utilized in the design and training of HMMs for the efficient classification of underwater targets. Simulation results are presented for typical underwater noise data waveforms.