Author(s): Matthew O. Ward, Tahar El-Korchi, Norman Wittels, Michael A. Gennert, Worcester Polytechnic Institute, Worcester, MA 01609

Source: Road and Airport Pavement Response Monitoring Systems, VC Janoo and RA Eaton, eds., pp41-62 (1992, ASCE, New York).

Abstract: Deterioration of the nations highway systems is occurring at an alarming rate, requiring maintenance costs in the billions of dollars annually. Because of the magnitude of the job, state and federal agencies are adapting and utilizing pavement management systems (PMS) to allocate funds more efficiently. Pavement surface conditions are important data used in determining the service adequacy of existing pavements. Current practices used to evaluate the severity and extent of pavement surface distress are laborious, subjective, and non-repeatable. Automation of pavement surface distress evaluation systems can provide high quality data that is objective and repeatable at a lower cost. However, this task has proven to be technically challenging.

This paper reviews current methods in image acquisition and image processing for automated pavement surface distress evaluation and presents recently developed engineering techniques for engineering and testing pavement image processing systems, an important step in overcoming the technical challenges inherent in designing a system capable of measuring the severity and extent of distress with features as small as 1.5mm (1/16") from a survey vehicle travelling at normal highway speeds.

This paper briefly describes how computer modeling of the imaging process can be used to design the image acquisition system, the video and computer hardware which generate and store the pavement images, and to optimize the lighting, the image processing algorithms, and the operation of the automated pavement distress evaluation system.

Finally this paper presents an engineering approach for testing the image processing portions of automated pavement surface distress evaluation systems. We show how a combination of naturally-acquired and computer-synthesized pavement images is optimal for system testing and present samples of system test images.