Video-based Seismology & Intelligent Earthquake Monitoring
video seismologyvirtual seismographanimal behavior recognitionlarge language modelsearthquake prediction
Seismic monitoring has long depended on dedicated instruments, creating a tension between coverage density and deployment cost. The lab explores cameras — “sensors that are already everywhere” — as a new means for earthquake monitoring.
What we do
- Virtual seismographs from video: video-based virtual seismograph methods and a fully-automated machine-vision smart seismograph (IEEE GRSL, Sensors); the ongoing project “AI-based video seismology models and engineering deployment” funded by the Institute of Geophysics, CEA.
- Intelligent animal behavior observation: an AI-based macro animal behavior observation and analysis system (Liaoning Earthquake Administration), with a survey published in Chinese Journal of Geophysics.
- LLM-driven strong-motion monitoring: a key S&T program of the Ministry of Emergency Management on LLM-driven intelligent strong-motion monitoring equipment and systems.
- Damage & casualty prediction: video-based building damage prediction and Bayesian machine-learning earthquake casualty prediction.
Selected publications
- Video-Based Method for Virtual Seismograph. IEEE GRSL, 2023.
- A Fully-Automated Smart Seismograph Based on Machine Vision. Sensors, 2019.
- Research on Machine Learning Methods for Earthquake Prediction Based on Bayes Theory. Pure and Applied Geophysics, 2025.