To Automate Detecting, Quantifying and Mapping of Delamination of Bridge Decks using Aerial Thermographic NDE

Background Timely detection of delamination/debonding issues of concrete bridge decks is essential to the safety of transportation and to make informed maintenance decisions. Many NDE methods (such as GPR, impact echo, chain dragging, half-cell, acoustic, and thermography) have been developed to identify these subsurface defects. Each of the developed methods has its pros and cons in terms of cost, safety, accuracy, reliability and operability. Most of these methods need to have traffic closure when conducting the detections, which can be a substantial burden and raise safety concerns for bridge users and DOTs. Therefore, delamination/debonding detection cannot be performed as often as necessary in order to have the most updated deck data for optimal decision making. Mixed results were reported by state DOTs using earlier thermographic detection methods, which relied on hand-held or vehicle-mounted thermal cameras for data collection of deck surface temperatures. For example, Iowa DOT did a series of studies from 1982 to 2013 on using thermal images to detect pavement/deck delamination. Notably, almost all these projects used raw thermal images to detect delamination. None of them used any image processing technologies. Major challenges in these earlier methods were high-noise, low-resolution/distorted thermal images and the lack of proper image processing capability. The outcomes of these methods were highly susceptible to the environmental noises such as shadows and surface textures. These challenges hindered the further adoption of the thermographic method as a viable detecting approach. The recent technical advancements in image processing and UAV technology has the potential to transform the thermographic-based approach to detect delamination/debonding/deterioration in concrete decks. The three most relevant new technical advancements include: (1) new algorithms in computer vision to detect edges and recognize objects in high-noise situations; (2) new image processing techniques for distortion correction and image stitching for creating seamless and accurate full deck images; and (3) high-resolution thermal and optical cameras mounted on UAVs to take close-up aerial thermal and optical images of the bridge decks to capture the tiny temperature variations of the deck surfaces. As a result, the revitalization of research activities in thermographic detecting has been witnessed recently with improved outcomes. The Pl proposes to investigate the performance of the new aerial high-resolution thermographic method in detecting, quantifying, and mapping delamination of bridge decks. The research team possess substantial expertise and research equipment in all the proposed technical areas, which represents a unique strength to tackle the challenges in NOE of delamination. Several experimental studies were conducted to develop appropriate image processing algorithms to get the delamination signals out of the surface noises caused by surface texture, shadow, dirt, gravels, etc. The preliminary results indicated that the detectability depth of the proposed thermography methods can reach: (1) 3.75 inches deep for delamination larger than 1 square foot; (2) 2.5 inches deep for delamination larger than % square foot. The research team tested the developed method in NDOT's two phases of the SHRP2 R06A project as extra references for comparison with BDl's results. The result of the delamination detection on Bridge-80770 was within expectation. A surprising finding came from the inspection of Bridge S006_30574. From the coring results, we found a strong tie between thermal features and higher chloride penetration/concentration in the deck. Since this finding came from a single case, more investigations are needed to confirm its validity. If confirmed, it will open a new avenue to determine the rates of deck deterioration and monitor the formation of the chloride concentration. The preliminary detection results of two other bridge decks: a concrete deck without overlay and a bridge deck with asphalt overlay indicate both decks have delamination/deterioration areas. Coring validations are needed to confirm the findings. Objectives There are three primary objectives in this project: 1) to further validate the proposed airborne thermographic approach in detecting, quantifying, and mapping delamination in bridge decks by substantially expanding the pool of candidate bridge decks in the project, including asphalt decks. The plan is to collect thermographic and optical image data of 30 candidate bridges identified by NDOT. The primary targets are bridges with concrete decks, including decks with concrete overlay and decks without concrete overlay. Per the TAC suggestion, we will also include some candidate bridge decks with asphalt overlay. 2) to identify the best timing and weather conditions for therrnographic surveying through numerical simulations. Numerical simulations will be conducted to identify the best timings to conduct thermographic bridge deck surveys across seasons in the Nebraska region by using boundary condition data from local weather stations. The results of the numerical simulations will be validated by outdoor lab experiments on the University of Nebraska-Lincoln campus. 3) to evaluate the feasibility and performance of machine learning methods in automat for detection, quantification, and mapping of concrete deck delamination .


    • English


    • Status: Completed
    • Funding: $109,844.00
    • Sponsor Organizations:

      Nebraska Department of Transportation

      1500 Nebraska 2
      Lincoln, NE  United States  68502
    • Project Managers:

      Halsey, Lieska

    • Performing Organizations:

      University of Nebraska, Lincoln

      1400 R Street
      Lincoln, NE  United States  68588
    • Principal Investigators:

      Shen, Zhigang

    • Start Date: 20190701
    • Expected Completion Date: 20201231
    • Actual Completion Date: 20201231
    • USDOT Program: Transportation, Planning, Research, and Development

    Subject/Index Terms

    Filing Info

    • Accession Number: 01705864
    • Record Type: Research project
    • Source Agency: Nebraska Department of Transportation
    • Files: RiP, STATEDOT
    • Created Date: May 24 2019 1:27PM