Detection Algorithm of Trapped Personnel Based on Sound Processing Technology

In the China's major disasters, the equipment which is used to rescue trapped personnel has high cost, needs to carry and has low efficiency. In order to solve the problems, a detection algorithm of trapped personnel based on sound processing technology (DATP) is proposed. In the DATP algorithm, sound signal is collected. The data with large energy is selected. Preprocessing method is used, such as Kalman filter and normalization. FFT transform is used to frequency domain analysis. The data is extracted in the band of 18400-19240 Hz and 20811-22050 Hz. The corresponding similarity is calculated. According to the similarity values, whether there is trapped personnel is determined. The simulation results show that DATP algorithm improves recognition rate and reduces false judgment rate and computation time. In short, in the sound recognition of trapped personnel when major disaster occurs, DATP is better than MFCC algorithm, and is suitable for embedded devices.