人民长江 ›› 2021, Vol. 52 ›› Issue (8): 198-203.doi: 10.16232/j.cnki.1001-4179.2021.08.030

• • 上一篇    下一篇

基于声谱图特征的硬岩破裂类型识别方法

黄杰,苏国韶, 王拯扶,刘岩鑫   

  • 发布日期:2021-09-27

Recognition method of hard rock cracking types based on spectrograms characteristics

HUANG Jie, SU Guoshao, WANG Zhengfu, LIU Yanxin   

  • Published:2021-09-27

摘要: 为揭示岩石破裂机理,提出了一种通过岩石破裂声音声谱图特征辨识岩石拉/剪破裂类型的方法。从巴西圆盘劈裂试验和直剪试验中分别收集具有代表性的张拉破裂和剪切破裂的声音信号,将声音信号绘制成声谱图,采用局部二值特征指标和颜色矩指标作为声谱图特征指标,进而建立了基于高斯过程分类的拉/剪声谱图自动识别方法。将该方法应用于双轴压缩试验中花岗岩拉/剪破裂过程识别,验证表明此方法具有较好的准确性和经济实用性。

关键词: 硬岩破裂; 声音信号; 声谱图形; 高斯过程分类;

Abstract: To further reveal the mechanism of rock cracking, a tensile/shear crack recognition method based on the characteristics of spectrograms was proposed.The representative tensile and shear sound signals were collected from the Brazilian disc splitting test and the direct shear test respectively.Then the sound signals were drawn into spectrograms.The local binary pattern indicator and the color moment indicator were used as characteristic index to distinguish spectrograms.Thus a tensile/shear crack recognition method based on Gaussian process classification was established.The method was applied to investigate the tensile and shear crack development of granite during biaxial compression test, proving that the proposed method was applicable and convenient.

Key words: hard rock cracking; sound signal; spectrograms recognition; Gaussian process classification;