【摘要】 层位特征是探地雷达路面检测的重要信息，而目前基于人工或相关算法的层位拾取方法存在主观性强、工作量大和每次仅能追踪一个层位等问题。为此，提出了一种基于探地雷达瞬时相位余弦的多层位自动追踪方法。首先，通过复信号分析获取了雷达数据的瞬时相位余弦；其次，利用子波余弦矩阵数据进行相似度分析后再计算其瞬时相位余弦，增强相位数据同相轴的横向连续性；然后，获取相位数据的空间位置、振幅和极性信息，并在信号幅值和同相轴特征等一系列约束条件下自动追踪横向连续的层位线；最后，通过比较深度方向相邻层位线上振幅的均方根平方值来确定层位数据及其极性，并通过设置层位线阈值和振幅阈值来提取强振幅连续的层位线数据。数值模拟和现场案例分析验证了本文方法的有效性和适应性。更多还原

【Abstract】 Horizon characteristics are important information in pavement detection using ground penetrating radar(GPR) data.However, current horizon picking methods based on manual work or related algorithms have problems such as strong subjectivity and heavy workload and they can only track one horizon each time.Therefore, this study proposed a multi-layer auto-tracking method based on the cosine of the instantaneous phase of GPR data.The specific steps of this method are as follows.Firstly, obtain the cosine of the instantaneous phase of GPR data through complex signal analysis.Secondly, carry out the correlation analysis of wavelet cosine matrix data and then calculate the cosine of the instantaneous phase of these data, aiming to enhance the transverse continuity of phase data along the cophase axis.Thirdly, obtain the spatial positions, amplitude, and polarity of the phase data, and automatically track the transversely continuous horizon lines under a series of constraints such as signal amplitude and cophase axis characteristics.Finally, determine the horizon data and their polarity by comparing the RMS values of the amplitude of adjacent horizon lines along the depth direction, and extract the horizon line data with continuous high amplitude by setting horizon and amplitude thresholds.Numerical simulation and field case analysis have verified the effectiveness and adaptability of the method proposed in this study.更多还原