Ocean Mammalian Sound Recognition Based on Feature Fusion
As development progresses, people become more aware of the importance of the environment. For marine mammals, environmental changes are likely to cause the extinction of the entire population, which is not conducive to the protection of marine mammals. Therefore, the effective search for marine mammals and their living areas is conducive to the protection of marine mammals. The use of the sound of marine mammals to identify and effectively judge the living area has become a hot topic in current research. Based on this, this paper also studies the marine mammalian voice recognition algorithm by extracting, merging and identifying features. Firstly, the LPCC (Linear Predictive Cepstral Coefficients) and MFCC (Mel Frequency Cepstral Coefficients) feature parameters and their extraction process are introduced in detail. Then based on the Extreme Learning Machine-Stack Self-Encoder (ELM-SAE) method, a method based on ELM-SAE feature fusion is proposed. Finally, the support vector machine is used to identify the features and realize the recognition of marine mammalian sounds. The simulation results show that the proposed feature fusion based marine mammal voice recognition algorithm has a good recognition rate, and the anti-jamming performance of background noise is also enhanced. It has a better application space in low signal-to-noise ratio marine environment.