Machine Learning Approaches for Road Condition Monitoring Using Synthetic Aperture Radar

Abstract

Airborne Synthetic Aperture Radar (SAR) has the potential to monitor remotely the road traffic infrastructure on a large scale. Of particular interest is the road surface roughness, which is an important road safety parameter. For this task, novel algorithms need to be developed. Machine learning approaches, such as Artificial Neural Networks (ANN) and Random Forest Regression, which can perform non-linear regression, can achieve this goal. This work considers fully polarimetric airborne radar datasets captured with DLR's airborne F-SAR radar system. Several machine learning-based approaches were tested on the datasets to estimate road surface roughness. The resulting models were then compared with ground truth surface roughness values and also with the semi-empirical surface roughness model studied in previous work.

Industry- and Academic-Based Trends in Pavement Roughness Inspection Technologies over the Past Five Decades: A Critical Review

“…The use of SAR technology for evaluating pavement roughness has been the focus of multiple studies [24,[129][130][131][132][133][134][135]. SAR data from X-band [132][133][134][135], L-band [24,131], and C-band [130] was investigated for pavement roughness estimation.…”

“…The use of SAR technology for evaluating pavement roughness has been the focus of multiple studies [24,[129][130][131][132][133][134][135]. SAR data from X-band [132][133][134][135], L-band [24,131], and C-band [130] was investigated for pavement roughness estimation. The data were collected using various satellites such as Sentinel-1 [24,130], Sentinel-2 [132], Cosmo-SkyMed [129], and Advanced Land Observing Satellite (ALOS) [131].…”

“…Thus, more research is required before assuming the robustness of this technology with the low-resolution publicly available satellite images. Alternatively, Babu et al [133] and Rischioni et al [135] used high-resolution airborne polarimetric SAR images to evaluate pavement roughness. The two studies utilized DLR's airborne radar sensor F-SAR to acquire fully polarimetric X-band datasets.…”

“…While satellite imagery is generally more cost-effective, airborne SAR images can offer higher resolution and greater accuracy. However, challenges such as the presence of vehicles on the road, tree coverage, and processing difficulties, including segmentation challenges, can limit the effectiveness of this approach in specific scenarios, particularly on urban roads [134,135]. Additionally, it was reported by Fiorentini et al [178] that roughness resulting from endogenous factors, such as traffic flow and pavement materials, may not be detectable using SAR technology.…”