Recognition of behavior state of Penaeus vannamei based on passive acoustic technology

Underwater signal processing based on passive acoustic technology has carried out a lot of research the behavioral sound underwater creatures and protection marine resources, which proves feasibility for long-term non-destructive monitoring biological production. However, at present, most relevant focuses fish but little shrimp. At same time, as main economic breeding industry, Penaeus vannamei backward industrial structure, in level intelligence needs to be improved. In this paper, signals generated by different physiological behaviors P. are collected technology. Their finally classified identified through feature extraction analysis. Meanwhile, characteristic non-parametric ANOVA is explore relationship between behavior state achieve purpose real-time . The experimental results show that linear prediction cepstrum coefficient (LPCC) Mel-frequency (MFCC) coefficients effective classification recognition with interspecific SVM classifier OvR strategy can model characteristics more efficiently accuracy high 93%.