Seismic signal detection and seismic phase extraction method

Abstract

The invention relates to a seismic signal detection and seismic phase extraction method which is used for a seismic detection system based on edge equipment. The edge device is realized by adopting a Jetson Nano chip, and a lightweight deep learning model LCANet is constructed and arranged on the edge device; seismic waveform data collected by the equipment end are inputted into the edge equipment, and a seismic signal time sequence, a longitudinal wave seismic phase and a transverse wave seismic phase are outputted in real time. Through the LCANet model, a feature vector sequence describing the intrinsic physical meaning of seismic data is extracted from input seismic waveform data through an encoder based on a reverse bottleneck residual block, and then a feature vector sequence of attention time sequence context information under three tasks is obtained through a context perception attention module; and finally, the feature vector is mapped to a feature space of a corresponding task through a multi-scale heterogeneous decoder. The model is very suitable for being deployed on edge equipment, and meets requirements of an earthquake early warning system for low delay and high performance of data processing.

Description

technical field [0001] The invention belongs to the technical field of earthquake early warning, and in particular relates to a method for seismic signal detection and seismic phase extraction. Background technique [0002] The earthquake early warning system is a large-scale Internet of Things system. According to the records of the National Earthquake Information Center, an average of 20,000 earthquakes occur every year around the world. Seismic stations around the world continuously generate data and transmit it to cloud servers, and seismologists analyze seismic data for earthquake research. Seismic data are widely used in earthquake location, earthquake early warning, earthquake signal detection, phase extraction, aftershock location prediction, etc. When an earthquake occurs, it releases energy in the form of seismic waves, which radiate in all directions from the source. Different types of energy waves shake the ground in different ways and travel through the Earth...

AI Technical Summary

Problems solved by technology

However, in the case of limited resources, the edge device usually only deploys a simple model-driven algorithm to the edge device, but its effect is poor and it is difficult to deal with complex real-world scenarios

However, data-driven deep learning algorithms contain complex network results and a large number of trainable parameters, which are more suitable for complex and changeable real-world scenarios. However, most deep learning algorithms cannot be applied due to their high computational complexity and large storage space. Resource Constrained Edge Devices

Images