A comprehensive study of the 2016 Mw 6.0 Italy earthquake based on high-rate (10 Hz) GPS data

The 2016 M_w 6.0 Italy earthquake is successfully recorded by the near-field 10?Hz GPS and 200?Hz Strong Motion (SM) stations, providing valuable data for this study. A comprehensive study of this earthquake is carried out based on GPS data, which contains coseismic deformations analysis, noise analysis, seismic wave picking, and magnitude determination. The noise of most GPS-derived displacement waveforms can be described as a combination of white noise, flicker noise, and random walk noise after the earthquake occurrence, and the spectral indices vary significantly for most stations, implying that the seismic signals have affected the noise characteristic of GPS-derived displacement waveforms. S-transform is employed to assess the GPS capability to detect the seismic arrival time. The SM station AMT and the GPS station AMAT are in good agreement in seismic wave picking, and the difference is only 1.2?s in the north component, suggesting that the outcome of seismic wave picking using GPS data is reliable. Then, a classic empirical formula is employed to determine the moment magnitude. A robust moment magnitude (Mw 5.90) can be estimated by the nine GPS stations with about 23.9?s. If four GPS stations near the epicenter is chosen to determine the magnitude, it only take 13.0?s to retrieve a reliable preliminary (Mw 5.82) magnitude, which is 5.4?s ahead of nine stations. In addition, Cross Wavelet Transform (XWT) is adopted to measuring the correlation and phase relationship between GPS and SM records. The result of XWT analysis indicates 10?Hz GPS is capable of capturing reliable and accurate coseismic dynamic deformations, as evidenced by the XWT-based semblance being close to 1 between GPS and SM records. The above results confirm the capability of 10?Hz GPS to capture coseismic dynamic deformations, detect seismic arrival time, and determine earthquake magnitude. Moreover, rapid magnitude determination based on 10?Hz GPS data can be regarded as an important supplement to Earthquake Early Warning (EEW).     

基于S变换的时频特征提取与目标识别

目标的时间-频率联合分布能够很好地反映目标物理结构特征,可以作为雷达目标识别的一个有效手段。针对现有时频分析方法存在的识别率低和抗噪性能差等问题,提出一种基于S变换的空间目标回波信号电磁特征提取与识别方法。首先对目标的雷达回波进行时频分析,得出在较大方位角变化范围内和不同信噪比情况下,目标S变换的时频分布具有相对不变性的结论;然后基于这种稳定的时频分布特征,采用最小贴近度的方法进行分类识别。针对不同目标模型的仿真结果表明,该方法的识别率高于其他时频分析方法并且具有很好的鲁棒性。     

基于广义S变换的图像局部时频分析

 非平稳信号具有良好的时频局部特性,但是使用一些常规的信号处理方法对其进行时频分析具有一定难度。为了解决这一难题,引入基于广义S变换的时频分析方法来进行一维和二维空间非平稳信号的时频分析。同时,为避开对图像直接进行S变换带来的大时间开销和运算难度,选用不同尺度的局部窗口将二维图像转化成一维信号进行处理,再分别利用基本S变换和时频分辨率可调的广义S变换进行算法仿真和时频分析。试验结果表明,广义S变换比基本S变换具有更灵活的时频聚焦性。