Graduation Year

2015

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Civil and Environmental Engineering

Major Professor

Manjriker Gunaratne, Ph.D.

Committee Member

Gangaram Ladde, Ph.D.

Committee Member

James Musselman, M.E.

Committee Member

Qing Lu, Ph.D.

Committee Member

Wilfrido Moreno, Ph.D.

Keywords

Asphalt Pavement, Human Visual System, Image Processing, Neural Network, Pie Plate Visual Method

Abstract

Florida Department of Transportation (FDOT) has been using Open Graded Friction Course (OGFC) mixture to improve skid resistance of asphalt pavements under wet weather. The OGFC mixture design strongly depends on the Optimum Binder Content (OBC) which represents if the mixture has sufficient bonding between the aggregate and asphalt binder. At present, the FDOT designs OGFC mixtures using a pie plate visual draindown method (FM 5-588). In this method, the OBC is determined based on visual inspection of the asphalt binder draindown (ABD) configuration of three OGFC samples placed on pie plates with pre-determined trial asphalt binder contents (AC). The inspection of the ABD configuration is performed by trained and experienced technicians who determine the OBC using perceptive interpolation or extrapolation based on the known AC of the above samples. In order to eliminate the human subjectivity involved in the current visual method, an automated method for quantifying the OBC of OGFC mixtures was developed using digital images of the pie plates and concepts of perceptual image coding and neural network (NN). Phase I of the project involved the FM-5-588 based OBC testing of OGFC mixture designs consisting of a large set of samples prepared from a variety of granitic and oolitic limestone aggregate sources used by FDOT. Then the digital images of the pie plates containing samples of the above mixtures were acquired using an imaging setup customized by FDOT. The correlation between relevant digital imaging parameters and the corresponding AC was investigated initially using conventional regression analysis. Phase II of the project involved the development of a perceptual image model using human perception metrics considered to be used in the OBC estimation. A General Regression Neural Network (GRNN) was used to uncover the nonlinear correlation between the selected parameters of pie plate images, the corresponding AC and the visually estimated OBC. GRNN was found to be the most viable method to deal with the multi-dimensional nature of the input test data set originating from each individual OGFC sample that contains AC and imaging parameter information from a set of three pie plates. GRNN was trained by 70% and tested by 30% of the database completed in Phase I. Phase III of the project involved the configuration of a quality control tool (QCT) for the aforementioned automated method to enhance its robustness and the likelihood of implementation by other agencies and contractors. QCT is developed using three quality control imaging parameters (QCIP), orientation, spatial distribution, and segregation of ABD configuration of pie plate specimens (PPS) images. Then, the above QCIP were evaluated from PPS images of a variety of independent mixture designs produced using the FDOT visual method. In general, this study found that the newly developed software (GRNN-based) provides satisfactory and reliable estimations of OBC. Furthermore, the statistical and computer-generated results indicated that the selected QCIP are adequate for the formulation of quality control criteria for PPS production. It is believed that the developed QCT will enhance the reliability of the automated OBC estimation image processing-based methodology.

Share

COinS