电磁波在岩体内的衰减和等距离岩体缺陷对同源反射波的叠加影响是探地雷达在地下岩体洞室围岩缺陷检测中影响雷达图像的两大难题，前者影响岩体缺陷大小的辨识，后者影响岩体缺陷位置的确定。虽然部分雷达数据采集软件对电磁波的衰减进行了一定的补偿，但同源反射波叠加问题则依赖于后期的软件处理及复杂的迁移计算，限制了探地雷达在围岩缺陷快速识别的应用和发展。同时，现有探地雷达成像方法无法清晰展现勘测结果，尤其当异状波形分布稀疏时，很难直观判断，限制了对探地雷达勘测剖面的解读。本文提出了适合探地雷达信号剖面快速成像的等值线算法，用于探地雷达成像。通过与现有成像软件Groundvision的结果对比，展示了成像结果的有效性。通过与等值色块图成像对比，并用MATLAB编制程序计算验证，展现了该等值线算法的高效性。该算法仅基于单一网格内的连接理论，未考虑复杂的拓扑关系，可以分区分块同时成像，是探地雷达成像的高效方法。本文基于麦克斯韦方程和均匀裂隙分布假设改进了波幅增益函数，用于补偿衰减的电磁波信号。将某地下管线勘测剖面和某地下洞室围岩勘测剖面，分别与现有软件Groundvision中指数增益函数和自动增益的成像结果进行对比分析，验证了本文波幅增益函数的有效性。且结果显示裂隙会造成电磁波强烈的衰减，在实际处理过程中是不可忽略的。此外，本文基于探地雷达的基本勘测理论，提出了不同于现有偏移方法的迭代消除方法，用于解决同源反射波的叠加问题。通过处理探地雷达勘测数据，并与Kirchhoff偏移方法对比，验证了本文迭代消除方法的有效性，并通过研究理论，对勘测区域数据分解至单一测线，可以得到用于实时传输和处理数据的迭代消除方法。最后，本文利用地下预埋管线勘测数据进行分析，验证本文方法的正确性。处理数据成像后，本文方法能够有效消除管线给周围区域造成的干扰信号，更加精确地显示管线的形状和位置，因而本文提出的处理探地雷达数据的方法是一种有效的方法。但本文研究方法理论存在一些假设和试算，这将有待进一步的研究。
The attenuation of electromagnetic waves and the superposition of reflected waves are two major issues that affect ground penetrating radar (GPR) imaging during the detection of surrounding rock defects in underground caverns, and influence the identification of rock defect size and location, respectively. Some GPR data acquisition software can compensate for the attenuated signals, but distinguishing true signals from the superposed reflected waves is dependent on subsequent artificial processing and complex migration calculations, which limits the application of GPR and related technologies. At the same time, the existing ground penetrating radar imaging methods cannot clearly show the survey results, especially when the irregular waveforms are sparsely distributed, which limits the interpretation of the ground penetrating radar survey profile.Based on the basic graphic theory, this paper proposes a contouring algorithm suitable for fast imaging of ground-penetrating radar signal profiles. Compared with the results of the existing imaging software Groundvision, the accuracy and clarity of the imaging results are demonstrated. Comparing with the equivalent color block diagram imaging, and using MATLAB to compile the program to calculate and verify, it shows the high efficiency of the contour algorithm. This algorithm is only based on the connection theory in a single grid, without considering the complex topological relations, and it can be applied for partitioned imaging simultaneously. In this paper, the amplitude gain function is improved based on the Maxwell's equation and the assumption of uniform crack distribution, and it is used to compensate for the attenuated electromagnetic signals. Using an underground pipeline survey profile and the surrounding rock survey profile of an underground cavern, the results were compared with the imaging results of the exponential gain function and automatic gain in the existing software Groundvision, respectively, to verify the validity of the amplitude gain function of this paper. Fissures will cause strong attenuation of electromagnetic waves, which cannot be ignored in the actual processing. At the same time, based on the basic survey theory of ground penetrating radar, this paper proposes an iterative elimination method that is different from the migration method to solve the problem of superposition of homogenous reflected waves. By processing the ground penetrating radar survey data and comparing it with the Kirchhoff migration method, the effectiveness of the iterative elimination method is verified. Through theoretical research, a single line decomposition of the survey area data can be used to obtain an iterative elimination method for real-time transmission and processing of data.Finally, the paper uses the survey data of buried pipelines for analysis to verify the correctness of this method. After processing data imaging, this method can effectively eliminate the interference signal caused by the pipeline to the surrounding area and display the shape and position of the pipeline more accurately. Therefore, the method proposed in this paper for processing ground penetrating radar data is an effective method. However, there are some assumptions and trial calculations in this research, which need to be further studied.
