Scholarlink

<p>In microseismic monitoring, the precision of phase arrival picking directly governs the reliability of event localization and focal mechanism inversion. In high-interference environments like coal mines, single-component array signals are often degraded by low signal-to-noise ratios and the combined presence of narrow-band mechanical and impulsive noise. Consequently, traditional picking methods based on time-domain energy or statistical shifts are prone to omissions and false triggers. This paper presents a collaborative phase-picking method for single-component arrays. First, station records are transformed into the time-frequency domain to generate instantaneous spectrograms. A frequency-wise normalization algorithm, based on the median and median absolute deviation (MAD), is implemented to suppress narrow-band common-mode interference. Subsequently, using local spectrogram blocks as templates, two-dimensional normalized cross-correlation (2D-NCC) is computed across stations in the time-shift dimension. An event detection function is then constructed by stacking these multi-station correlation responses. Leveraging the spectral differences between P and S phases, high- and low-frequency correlation stack curves are established to facilitate a sequentially constrained “detection–P picking–S picking” sequence. Additionally, confidence metrics are derived from stack peaks and inter-station correlation consistency. Simulation results involving colored noise, narrow-band interference, and impulsive noise demonstrate that this method achieves stable event detection under low signal-to-noise conditions. It significantly improves the temporal consistency and interference rejection of P/S picking. Frequency-wise normalization and PSR-like saliency measures further enhance peak discriminability and reduce threshold sensitivity. These findings indicate that the array-based collaborative processing scheme provides a logically rigorous and highly stable technical pathway for automated single-component microseismic picking.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Research on Collaborative Microseismic Phase Picking for Single-Component Arrays Based on Time-Frequency Two-Dimensional Cross-Correlation

  • 期刊论文

  • 发表日期：

    2026年6月17日

查看全文
  • Gui-wu Chen,
  • Yan Li,
  • Shu-xun Zhang,
  • Hao Wang,
  • Shao-lei Song,
  • Peng-Cheng Huang

摘要

In microseismic monitoring, the precision of phase arrival picking directly governs the reliability of event localization and focal mechanism inversion. In high-interference environments like coal mines, single-component array signals are often degraded by low signal-to-noise ratios and the combined presence of narrow-band mechanical and impulsive noise. Consequently, traditional picking methods based on time-domain energy or statistical shifts are prone to omissions and false triggers. This paper presents a collaborative phase-picking method for single-component arrays. First, station records are transformed into the time-frequency domain to generate instantaneous spectrograms. A frequency-wise normalization algorithm, based on the median and median absolute deviation (MAD), is implemented to suppress narrow-band common-mode interference. Subsequently, using local spectrogram blocks as templates, two-dimensional normalized cross-correlation (2D-NCC) is computed across stations in the time-shift dimension. An event detection function is then constructed by stacking these multi-station correlation responses. Leveraging the spectral differences between P and S phases, high- and low-frequency correlation stack curves are established to facilitate a sequentially constrained “detection–P picking–S picking” sequence. Additionally, confidence metrics are derived from stack peaks and inter-station correlation consistency. Simulation results involving colored noise, narrow-band interference, and impulsive noise demonstrate that this method achieves stable event detection under low signal-to-noise conditions. It significantly improves the temporal consistency and interference rejection of P/S picking. Frequency-wise normalization and PSR-like saliency measures further enhance peak discriminability and reduce threshold sensitivity. These findings indicate that the array-based collaborative processing scheme provides a logically rigorous and highly stable technical pathway for automated single-component microseismic picking.