Time Frequency Domain Analysis of Internal Diseases in Asphalt Pavement Based on Ground Penetrating Radar

Abstract

Ground penetrating radar (GPR), often referred to as the “CT” of road, can be applied to conduct a comprehensive analysis of the internal structure and potential defects of roads, thereby providing an accurate assessment of its overall condition. However, current qualitative analysis of the collected radar data primarily relies on mapping, which limits the ability to effectively assess the varying degrees of damage caused by the underlying defects. In this paper, gprMax is used to model the forward simulation of GPR signals through asphalt pavement, focusing on cracks, interlayer defects, loose structures. The study explores the time-frequency characteristics of varying degrees of damage caused by these defects, aiming to analyze how the time-frequency eigenvalues change as the severity of the damage increases. A sensitivity analysis is conducted to examine the sensitivity of different time-frequency eigenvalues to variations in the degree of damage. The non-stationary characteristics of signals along both the time and frequency axes are analyzed using wavelet transform. Finally, the results of the forward simulation are validated against the measured data. The results show that five time-domain eigenvalues and four frequency-domain eigenvalues are effective in comprehensively characterizing varying degrees of damage caused by three types of defects: cracks, interlayer defects, and loose structures. Specifically, the trend of change in crack eigenvalues is opposite to that of interlayer defects and loose structures. Among the time-frequency eigenvalues, the square difference (SQD) and quadratic variance (QV) in the time domain, as well as the mean value (MV) and standard deviation (STD) in the frequency domain, exhibit greater sensitivity to the degree of road damage. As the degree of damage increases, the shifts in peak frequency for three defects differ. The peak frequencies of cracks, at 1.4 and 2.0 GHz, shift to higher frequencies, while the peak frequency of interlayer defects, at 2.0 GHz, shifts to lower frequencies. In contrast, the peak frequency of loose structures, at 1.4 GHz, shifts to higher frequencies. In summary, time-frequency characteristics offer an effective approach to analyzing varying degrees of damage caused by the internal road defects, thereby providing theoretical support for the advancement of ground penetrating radar detection from qualitative image recognition to quantitative analysis.