Date of Award
Winter 2024
Project Type
Dissertation
Program or Major
Civil and Environmental Engineering
Degree Name
Doctor of Philosophy
First Advisor
Eshan E.V.D. Dave
Second Advisor
Fei Han
Third Advisor
Yaw Adu-Gyamfi
Abstract
This research leverages Unpiloted Aerial Systems (UAS) equipped with advanced sensing technologies—optical imaging and Light Detection and Ranging (LiDAR)—to accurately detect and measure pavement frost heaves, which cause significant structural damage and surface roughness, particularly in cold climate regions. The use of UAS platforms is driven by the need to provide measurements over large areas and to offer high-precision, accessible, and efficient data collection methods capable of operating in challenging and difficult-to-access regions. Furthermore, UAS technology enhances safety and saves time and labor, enabling quicker, and safer assessments with minimal disruption. Experiments were conducted on both simulated and actual pavements that provided insights into optimal flight parameters. For photogrammetry, flying at an altitude of 50 m at a speed of 5 m/s with image captures every two seconds using a Sony a5100 camera with a 24 megapixels sensor resulted in substantial image overlap, aiding the construction of detailed pavement surface profiles from the data. The surface roughness was quantified using the International Roughness Index (IRI), which showed an increase with the onset of colder seasons, corroborated by comparisons with instrumented vehicle data. Concurrently, LiDAR-based sensing approaches were evaluated, determining that a lower flight altitude of 45 m at a ground speed of 2 m/s, combined with a 10 cm×10 cm spatial resolution, optimized the measurement of frost heave induced pavement surface elevation changes. This refined flight protocol and data processing method were then applied to in-service pavement sites across multiple seasons.This study applies second derivative analysis and distribution plots with overlapping window analysis on road segments to monitor pavement conditions over time and through seasonal changes. It uses high-resolution LiDAR data on a 10 cm×10 cm grid for a comprehensive topographical analysis, which includes cleaning data to remove anomalies, smoothing elevation profiles, and fitting polynomial curves. Residuals from these curves help identify and quantify maximum vertical deviations. Histograms illustrate shifts in pavement roughness, particularly noting increased roughness in colder months, indicative of potential frost heaves. Overall, this approach assesses pavement smoothness and identifies significant vertical deviations.
Recommended Citation
Zaremotekhases, Farah, "UAS-BASED DETECTION AND QUANTIFICATION OF PAVEMENT ROUGHNESS AND VERTICAL DEVIATIONS DUE TO FROST HEAVE" (2024). Doctoral Dissertations. 2885.
https://scholars.unh.edu/dissertation/2885