As one of the most earthquake-prone regions in the world, China faces extremely serious earthquake threats, especially for those heavily populated urban areas located near large fault zones. To improve the ability to prevent and minimize earthquake disaster risks, and to reduce earthquake disaster losses, China is currently building a nationwide earthquake early warning system (EEWS) with the largest seismic network in the world. In this paper, we present the newest progress of this project through describing the overall architecture of the national EEWS and evaluating the system performance during the 2022 Lushan M6.1 earthquake. The accuracy of the source characterization for the Lushan earthquake is discussed by comparing the continually estimated location and magnitude with the catalogs obtained from the China Earthquake Networks Center. For this earthquake, the EEWS generated a total of five alerts, and an initial alert was created 5.7 s after its occurrence, with excellent epicentral location and origin time estimation. The final alert was issued 16.5 s after origin time with a magnitude estimate of M6.1, the same as the catalog value. However, from the point view of alerting performance, the radius of the real blind zone without warning time was about 30 km and much larger than the theoretical result, mainly caused by the releasing system not considering the epicenter distance of each terminal when issuing the alerts. Although the earthquake exposed some limitations that need to be addressed in future upgrades, the results showed that most aspects of the EEWS presented a robust performance, with continuous, reliable event detections and early-warning information releasing.

Rapid seismic intensity estimation supports situational awareness, emergency response, and the assessment of earthquake-induced structural risk. Instrument-based systems provide authoritative measurements, but limited deployment density and installation cost can restrict spatial coverage. Video sensing offers a complementary route by converting the visible motion of a calibrated target into physical motion indicators; its reliability, however, depends on stable frame-wise localisation under illumination variation, blur, and target deformation. This paper presents FAMTrack, a vision-based seismic intensity prediction framework whose tracking module augments a one-stream Vision Transformer with a parallel Frequency-Aware Matching branch that encodes patch-wise Fourier amplitude and phase. A Cross-Domain Fusion Module exchanges spatial and frequency evidence through bidirectional cross-attention without modifying the detection head or tracking loss. On LaSOT, FAMTrack-256 attains a 75.1% AUC, 4.0 percentage points above OSTrack-256. On GOT-10K, it reaches 77.9% AO, a gain of 4.2 points. In the held-out seismic-video evaluation, FAMTrack reduces the displacement RMSE from 0.523 to 0.332 mm and improves window-level intensity prediction accuracy from 80.0% to 86.7% relative to OSTrack-256. Under the fixed calibrated-video pipeline and the evaluated controlled conditions, these results indicate that frequency-aware localisation is associated with improved public-benchmark tracking, motion recovery, and seismic intensity estimation.