Automatic Road Subsurface Distress Recognition from Ground Penetrating Radar Images using Deep Learning-based Cross-verification

TL;DR

This paper introduces a deep learning-based cross-verification method using YOLO models to automatically detect road subsurface distress from GPR images, achieving high accuracy and significantly reducing manual inspection effort.

Contribution

A novel cross-verification strategy leveraging multiple YOLO models for improved RSD detection from GPR images, enhancing accuracy and efficiency.

Findings

Recall over 98.6% in field tests

Automatic detection reduces human labor by around 90%

Effective recognition of voids and loose structures in GPR data

Abstract

Ground penetrating radar (GPR) has become a rapid and non-destructive solution for road subsurface distress (RSD) detection. However, recognizing RSD from GPR images is labor-intensive and heavily relies on the expertise of inspectors. Deep learning-based automatic RSD recognition, though ameliorating the burden of data processing, suffers from insufficient capability to recognize defects. In this study, a novel cross-verification strategy was proposed to fully exploit the complementary abilities of region proposal networks in object recognition from different views of GPR images. Following this strategy, three YOLO-based models were used to detect the RSD (voids and loose structures) and manholes. Each model was trained with a specific view of 3D GPR dataset, which contains rigorously validated 2134 samples of diverse types obtained through field scanning. The cross-verification strategy achieves outstanding accuracy with a recall of over 98.6% in the tests using real field-scanning data. Field tests also show that deep learning-based automatic RSD recognition can reduce the human labor of inspection by around 90%.