摘要
在背景条件复杂的工区,为了提高探地雷达(GPR)勘探资料解释的准确性和可靠性,利用独立分量分析理论进行了强噪声背景下的探地雷达信号去噪研究。阐述独立分量分析(ICA)基本理论,着重讨论了基于负熵最大化的快速独立分量分析(FastICA)算法。应用FastICA算法对单道探地雷达数据和正演含噪雷达剖面分别进行去噪分析,得到去噪后的探地雷达信号。以湖北恩施彭家寨隧道GPR实测数据为例,将Fast ICA算法应用于探地雷达剖面数据去噪。研究结果表明,将FastICA算法应用于探地雷达信号处理,摆脱了传统方法参数设置的束缚,流程简单,在GPR去噪方面有独特的优势,可较好地对低信噪比的GPR原始数据进行噪声去除,有助于突出探地雷达剖面中异常体特征,达到了提高资料解释准确性和可靠性的目的。
To improve the accuracy and reliability of interpretation of ground penetrating radar （GPR） complex area,independent component analysis （ICA） is used for GPR data de-noising process in strong background noise. In this work,the basis theory for ICA is firstly introduced,and negative entropy -based FastICA algorithm is discussed in de-tails. In the following discussion,applied to the de-noising process of single channel GPR datasectional data with noise. Besides,the GPR data from Hubti Enshi transit tunnel is collected for the further test of this FastI-CA algorithm in this paper. The results show that FastICA algorithm overcomes the constraints on paramethods for GPR data process,simply process,and has advantage on GPR data de - noising, especially for GRP data. The de -noising data can highlight the characteristics of anomalies,which is helpful for improving the accuracy and reliability of interpretation.
出处
《物探化探计算技术》
CAS
CSCD
2017年第6期727-735,共9页
Computing Techniques For Geophysical and Geochemical Exploration
基金
国家自然科学基金资助项目(41374118)
