Applied research of deep learning technology in the classification of earthquake and blasting event
Research Square (Research Square)(2023)
摘要
In order to quickly and accurately identify the types of natural earthquake and non-natural earthquake events, this paper proposes a lightweight convolutional neural network model. Since generally an event can be recorded by several stations, data preprocessing and classification based on the event shall be carried out in advance in order to avoid the occurrence of different station waveforms of the same event in any 2 of the training set, validation set and test set. With the three component waveforms recorded by stations after preprocessing as the input, the network model and hypoparameters are optimized by analyzing the average and variance of the accuracy and loss values of the verification set in the five-fold cross-validation, as well as the accuracy and loss curves in the training process. Finally, the classification results of all stations that achieve a certain signal-to-noise ratio for each event are taken as the output of this event type based on the principle that the minority is subordinate to the majority. In this study, 2190 natural earthquake and non-natural earthquake events recorded by the Hainan Seismic Network before August 2022 which contains 53067 waveforms are used to train and test the effect of the model. 20% of those events are selected randomly as the test set. The results showed that 427 of 438 randomly selected events were correctly identified, witch means that the accuracy rate is 97.48%.Among them, the accuracy rate of seismic events was 95.59%, the recall rate was 89.04%, the accuracy rate of blasting events was 97.84%, and the recall rate was 99.18%. To sum up, the convolution neural network model constructed in this paper can quickly and accurately identify the types of natural and non-natural earthquakes in Hainan.
更多查看译文
关键词
deep learning technology,deep learning,earthquake,classification
AI 理解论文
溯源树
样例
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要