Thermal Segregation and Compaction Analysis of HMA Construction using Unmanned Aerial Vehicles

Abstract: Thermal uniformity of hot mix asphalt paved mat is essential to achieve target density and designed performance. Existing techniques like Paver Mounted Thermal Profilers and Intelligent Compaction Technologies quantify thermal non-uniformities and density differentials. This study develops a protocol to improve construction quality using unmanned aerial vehicle (UAV) assisted aerial thermal images. Thermal images are collected from various sites visited in 2023 and 2024. Pavement sections were instrumented with thermocouples to validate the thermal reading from the aerial images.
A new technique called “Mat Differential Matrix” (MDM) was developed to quantify non-uniform patterns as a function of time. The MDM results are compared with the existing quantification metrics for thermal segregation, including Differential Range Statistic (DRS) and Thermal Segregation Index (TSI). The results showed that MDM agrees with DRS and TSI in most cases. All three methods assessed the thermal uniformity of the paved mat. However, MDM offered an advantage by determining the regions with higher accuracy and granularity of thermal segregation.
Aerial thermal images were utilized to establish a framework to track rollers’ movements and compaction uniformity. A deep learning computer vision model is developed to achieve this objective. The model is trained on data from 15+ sites by tracking the roller movements and pavement coordinates within the frames at a 2-sec interval. The developed protocol showcases the automated and implementable use of thermal images captured from the UAVs, providing the contractor with spatial coverage and near real-time feedback.

Keywords: Unmanned Aerial Vehicle, Thermal Imaging, Thermal Segregation, Non-uniform Temperatures, Object Detection, Roller Compaction, Thermal Metrics, Object Tracking