Sound signal feature extraction of mine roof percussion based on human auditory model

Human auditory model was introduced, vibration model of the basement membrane, inner hair cells model and cochlear nucleus mathematical model were analyzed, and extraction process of auditory spectrum feature vector was given. On the above basis, sound signal feature extraction method of mine roof percussion based on human auditory model was proposed. Human auditory model and wavelet packet were used respectively for sound signal feature extraction of mine roof percussion, and then support vector machine classifier was used for target feature classification and recognition. The experimental results show that correct identification rate of the feature extracted using human auditory model is above 95%, which indicates the sound signal feature extraction method of mine roof percussion based on human auditory model will help improve the accuracy of coal mine roof detection.