New Method for Classifying and Quantifying Cracking of Flexible Pavements in Automated Pavement Condition Survey

Abstract

The automated pavement condition survey is gradually replacing traditional visual surveys in many highway agencies. Incompatibilities between existing survey manuals and computer imaging technologies must be addressed to obtain maximum benefits from the new survey method. Existing survey manuals are based on the ability of human surveyors to subjectively recognize distresses by identifying crack patterns in various locations on the pavement surface. New imaging technologies provide grayscale images with precise location identification and therefore computer algorithms can be used with the imaging technologies to aid and (currently) partially replace the human. A newly developed method for the classification and quantification of flexible pavement cracks is being implemented by the California Department of Transportation to take maximum advantage of new technologies while strengthening the application of the pavement condition survey in separating cracks caused by different mechanisms. Cracks are categorized into four types by mechanism and quantified by geometrical features in images, including locations, orientations, and lengths. Subjective assessments regarding interconnectivity, pattern, and causes are eliminated. In the redesigned data flow, the automated pavement condition survey service providers perform image collection and analysis and submit objective measurements regarding geometrical features of cracking at the pavement data segment level. It is the highway agency's task to segment the highway network into management segments and derive summary distress quantifications at the management segment level on the basis of the data segment level survey results. This arrangement enables service providers to handle objective measurements and the associated constant technological innovation and gives the agency the flexibility to apply pavement engineering and management principles.