A Deep Learning Tool for the Assessment of Pavement Smoothness and Aggregate Segregation during Construction
Research studies indicate that Non-Destructive Testing (NDT) methods have the potential for use in the Quality Assurance (QA) of pavement construction since they allow for (a) fast evaluation of the product uniformity in real time as construction progresses; (b) identifying potential defects during construction to allow for timely corrective actions; (c) more frequent inspecting, testing, and replicating without the damaging effects of coring and other destructive testing; and (d) reducing the testing and inspections costs, while improving construction quality. For example, results of NCHRP 10-65 showed that GeoGauge is the device recommended for estimating the modulus of unbound layers, while the portable seismic pavement analyzer (PSPA) was recommended for estimating the modulus of Hot Mix Asphalt (HMA) layers. However, in spite of their high potential and usefulness, the transition of NDT methods from research to QA programs has been limited, and the destructive and time-consuming process of coring and laboratory testing continues to be the most widely used QA methods in the US. Pavement roughness, described in terms of the International Roughness Index (IRI), is the most widely used index for pavement condition evaluation; its use in quality assurance has also increased in recent years. Roughness is the opposite of pavement smoothness, which is a measure of ride quality and driver’s comfort. IRI represents the deviation of the pavement surface from the leveled plan that affects vehicle movement and ride quality. Pavement smoothness is measured using a profilometer or a laser-based surface tester. Technically, IRI is the cumulative vertical displacement of an axle from a reference quarter-car divided by the distance traveled over the pavement profile at a standard speed of 50 mph. Many new specifications require all mainline paving meet surface profile smoothness tolerance using IRI for quality assurance requirements. Including surface profile smoothness tolerance requirements in specifications also allows for an additional payment incentive or penalty based on the contractor’s performance. While the use of smoothness specifications is a positive development in pavement engineering to assess construction quality, the unavailability of a profiler in most of the road construction projects is a major obstacle for widespread implementation of this practice. Hence, the measurement of pavement smoothness and the detection of aggregate segregation in realtime or upon completion of the construction process is rarely conducted in practice especially in secondary roads. This may result in inadequate construction quality and the inability to introduce timely remedies to address the noted deficiencies. The present study will develop a machine learning-based classifier for the prediction of pavement smoothness and aggregate segregation based on digital image analysis and deep learning models. The proposed classifier will be developed such that it can be used by site engineers to predict pavement smoothness and aggregate segregation in real-time using camera-captured pavement images during the construction phase. The developed classifier and application may be used by the site engineer during construction activities to assess the smoothness of the paving mat and to detect the presence of segregation. This tool will be practical and straightforward and may be used by state agencies in case of the inaccessibility of laser-based surface tester or profilometer.
- Record URL:
-
Supplemental Notes:
- 21COLSU14
Language
- English
Project
- Status: Active
- Funding: $60000
-
Contract Numbers:
69A3551747106
-
Sponsor Organizations:
Office of the Assistant Secretary for Research and Technology
University Transportation Centers Program
Department of Transportation
Washington, DC United States 20590 -
Managing Organizations:
Transportation Consortium of South-Central States (Tran-SET)
Louisiana State University
Baton Rouge, LA United States 70803 -
Project Managers:
Mousa, Momen
-
Performing Organizations:
Louisiana State University, Baton Rouge
P.O. Box 94245, Capitol Station
Baton Rouge, LA United States 70803 -
Principal Investigators:
Elseifi, Mostafa
- Start Date: 20210801
- Expected Completion Date: 20230201
- Actual Completion Date: 0
- USDOT Program: University Transportation Centers Program
Subject/Index Terms
- TRT Terms: Aggregates; Image analysis; Machine learning; Paving; Quality assurance; Smoothness
- Subject Areas: Construction; Highways; Materials; Pavements;
Filing Info
- Accession Number: 01787557
- Record Type: Research project
- Source Agency: Transportation Consortium of South-Central States (Tran-SET)
- Contract Numbers: 69A3551747106
- Files: UTC, RIP
- Created Date: Nov 10 2021 8:18AM