Evaluation of Errors in Automated Pavement‐Distress Data Acquisition

Abstract

The extent of degradation and decay of transportation infrastructure has created new demands on maintenance management systems. This, in turn, has emphasized the need for accurate condition assessment data for use in infrastructure management. At the same time, technological advances in automated inspection systems provide the opportunity to automate the collection of surface condition data. The effect of the technology on accuracy has not been previously evaluated. This paper evaluates the data‐acquisition component of two automated optical technologies and identifies the main sources of error. Simulated longitudinal and transverse cracks are used to illustrate the results. The simulations indicate that correction factors are important to compensate for errors due to resolution and coverage limitations. Therefore, on a crack‐by‐crack basis, most technologies will perform poorly. When cracks are uniformly distributed, equipment with high resolution but poor coverage provides more accurate estimates of cracking extent but has more variability than equipment with good coverage. Finally, the results indicate that the inherent variability in the location, width, and length of cracks may lead to significant error.