基于2D-LMS的探地雷达背景噪声去除

GPR background noise removal based on 2D-LMS

铁路建设中常需在铁路两侧构筑挡土墙对铁轨进行保护,因此挡土墙施工质量对铁路具有重要意义。为了获得更有效的挡土墙质检数据,常用探地雷达(ground penetrating radar,GPR)进行质检。结合理论分析了电磁波在地下非均匀介质边界的反射和衍射特性,结果表明回波噪声呈二维分布。对仿真回波二维加噪后,对比二维中值和一维自适应(1D-LMS)算法滤波效果,仿真结果表明二维处理明显优于一维。为适应GPR回波信号的非线性特征,选取二维自适应(2D-LMS)算法对回波信号进行处理,并与二维中值滤波相比较。广东省某铁路挡土墙实测数据处理结果表明,2D-LMS算法滤波信号具有更高信噪比,能更有效地去除噪声,使回波图像更加清晰。

Retaining walls are always needed on both sides in the railway construction to protect the rail, thus the construction quality of the retaining walls are very important. To obtain more eff-ective quality-inspection data of retaining walls, the GPR （Ground Penetrating Radar） is often sel-ected. The paper analyzed the reflection and diffraction characteristics of the electromagnetic wav-es in the underground inhomogeneous-medium boundary. The results show that the echo noise pr-esents the two-dimensional distribution. After adding two-dimensional noise to the simulation echo, and by the comparison between the two-dimensional median filtering and the one-dimensional ad-aptive （II）-LMS） algorithm filtering, the simulation results show that the two-dimensional treat-ment is apparently better than the one-dimensional treatment. To meet the non-linear characterris-tics of GPR echo signal, the paper selected the two-dimensional adaptive （2D-LMS） algorithm for the echo signal processing, and compared it with the two- dimensional median filtering. The results of tile measured data from the certain railway retaining walls in Guangdong Province showed that.under the 2D-LMS algorithm, the filtered signal was with a higher signal to noise ratio（SNR）, the noise could be removed more effectively, and the echo images were clearer.