To improve the accuracy and real-time transmission of earthquake warning data,a real-time high-frequency cloud receiving algorithm for earthquake warning data based on deep neural networks is designed.Using deep neural network algorithms to search for earthquake warning data,using zero analysis method to determine the correlation between different types of vibration frequencies,and using each cell in deep neural network to simulate seismic data,denoise the data,and identify fluctuation signals,this paper transforms classification signals through Fourier transform algorithm to achieve real-time high-frequency cloud recep-tion of earthquake warning data through deep neural networks.The experimental results show that the proposed algorithm has a receiving error of less than 5mgal for tapping vibra-tion data,a data acquisition time of less than 42.5 ms,an amplitude signal receiving error of less than 1 m·s-1,a speed fitting error of less than 0.1 km·s-1,and a received signal en-ergy deviation of less than 0.004 J.The above data prove that the receiving algorithm has high accuracy in receiving impact vibration,seismic amplitude,velocity fitting and energy value.It can retain the original seismic signal more completely,and improve the real-time and efficient cloud reception of earthquake warning data.
关键词
深度神经网络/地震预警数据/实时/高频云/接收/噪声
Key words
deep neural network/earthquake early warning data/real time/high frequency cloud/receive/noise
维普
万方数据