|
ИСТИНА |
Войти в систему Регистрация |
ИСТИНА ИНХС РАН |
||
Earthquake Arrival Identification in a Record with Technogenic Noiseстатья
Дата последнего поиска статьи во внешних источниках: 27 февраля 2020 г.
- Авторы: Kislov K.V., Gravirov V.V.
- Журнал: Seismic Instruments
- Том: 47
- Номер: 1
- Год издания: 2011
- Первая страница: 66
- Последняя страница: 79
- DOI: 10.3103/S0747923911010129
- Аннотация: These papers consider the problem of identifying a seismic event in increased noise based on data recorded by a single seismometer or a local array of seismometers. The problem is the key to the development of an early warning system about an earthquake that has occurred. Catastrophic damage and the loss of human lives usually result from the suddenness of earthquakes. If data processing and transmission of information can be carried out very fast (within 4-5 s), such an earthquake warning will be valuable as reducing the loss of human lives and the economic loss. The goal of this work is to study the selection of information features of a signal with a view to reducing their dimensionality and to learn about the peculiarities of self-contained systems. The emphasis was on the study of how neural networks can be used to examine class separability in the feature space. The system we propose is to reduce the negative impact of a damaging earthquake. We highlight the leading issues and discuss methods toward their solution; the testing was conducted using a trial problem: identification of small earthquakes in a noise-contaminated record. We state the main parameters of the relevant algorithms and programs for identification of onsets of earthquake phases in high industrial and other noise.
- Добавил в систему: Гравиров Валентин Валентинович