Non-destructive testing and intelligent evaluation of road structural conditions using GPR and FWD

Abstract

Aiming at the assessment of road structural conditions, a method of combining ground penetrating radar (GPR) and falling weight deflectometer (FWD) for assessing the pavement structural integrity and strength was proposed in this study. First, 3D GPR was performed to detect the thickness and internal distress of pavement structural layers using the canny edge detection and you only look once version eight (YOLOv8) detection algorithms. Results showed that the error of thickness extraction was approximately 3%, and the distress detection achieved a mean average precision (mAP) of 0.859 and an inference time of 11.43 ms with a GTX 1070 GPU. Then, the extracted thickness was used for the modulus (E) inversion of pavement structure layers based on FWD test data and regression analysis. Finally, the distress ratio inside pavement structures (DRIPS) was proposed as a structural integrity index. The relationship between E and DRIPS was revealed and showed a good correlation. The greater the DRIPS value, the worse the pavement structural integrity and strength. It provides a reference for the evaluation of road structural conditions. This strategy proved to be reliable for nondestructive testing and evaluation of road structures, which could improve the comprehensiveness and effectiveness for evaluating road structural conditions.