A Feature Extraction Method Based on Continuous Spectrum in Gammatone Frequency Bands

Target recognition is one of the most important technologies in the field of underwater acoustic detection, and has been widely used in the military. Target radiation noise and environmental noise are random signals, environmental noise has obvious time-varying characteristics, so a specific method is difficult to achieve multi-class similar target classification and recognition. For the targets with similar auditory characteristics, take first preliminary classification and then accurate recognition as technical ideas, a feature extraction method based on continuous spectrum in Gammatone frequency bands is proposed. Initial target classification is realized through Gammatone feature extraction, and accurate identification is realized through extracting Continuous spectrum feature in Gammatone subband. By using the real recording data of surface target, verifying the proposed method. Results show that the proposed method can effectively classify and recognize two kinds of objects with similar auditory characteristics.