ASSESSING THE FEASIBILITY OF UTILIZING A UAS-BASED POINT CLOUD IN PAVEMENT SMOOTHNESS/ROUGHNESS MEASUREMENT

Abstract

The metric derived from the longitudinal profile IRI (international Roughness Index) is a substantial input for the majority transportation agencies' highway monitoring systems used to improve road safety, increase road quality, and reduce fuel consumption. Traditionally, authorities have used data collected by inertial profilers to evaluate IRI; however, these instruments are usually unfeasible for data collection in smaller areas, and their narrow field-of-view (FOV) produces inadequate context for the scene. Meanwhile, uncrewed aircraft systems (UAS) have been widely used in a variety of transportation applications because of their efficiency and affordability in acquiring high-quality data. Especially for smaller areas, through the Structure from Motion (SfM) technique, UASs are a promising complementary tool for conducting ground surveys because they can provide good 3D context with high-resolution images. However, the SfM approach has some limitations for predicting the accuracy and quality of produced point clouds because of some factors such as surface texture, lighting conditions, processing algorithms/software etc. Our study aimed to assess the feasibility and data accuracy of the SfM technique for evaluating IRI by considering its limitations. This study also sought to build a framework for obtaining IRI metrics from an SfM-derived point cloud and to provide recommendations for collecting and processing UAS data with the goal of extracting pavement information.