Automated Identification and Quantification of Flexible and Rigid Pavement Distresses

This research project was conducted with the objective of evaluating the viability of using an automated computer application to detect, classify, and quantify cracks on flexible and rigid pavement surfaces from digital images, to be implemented in the annual Pavement Condition Survey (PCS). Phase I of this project was focused on identification and quantification of the rigid pavement cracks from two-dimensional (2D) images collected using the FDOT data collection vehicle. The developed FDOT Rigid Pavement Distress Application (FRPDA) was able to detect joints and assign slabs, detect, classify, and quantify longitudinal, transverse, and corner cracks with higher accuracy, repeatability, and efficiency compared to the current manual windshield method of the PCS. Phase II of this project was focused on identification and quantification of the flexible pavement cracks from three-dimensional (3D) images collected using the Fugro data collection vehicle. The developed FDOT Flexible Pavement Distress Application (FFPDA) was able to detect, classify, and quantify all crack types with higher accuracy, repeatability, and efficiency compared to the current manual windshield method of the PCS. It is recommended that the automated FRPDA and FFPDA be used on a longer length of highway pavements along with a semi-automated survey to conduct quality control of the results and to identify and quantify non-cracking distress types for the PCS. The report is available online at https://fdotwww.blob.core.windows.net/sitefinity/docs/default-source/research/reports/fdot-bdz36-rpt.pdf