A pavement condition survey is one of the most essential elements of any pavement management system. During the last decade, significant progress has been made towards automating pavement distress survey procedures by use of advanced imaging technology without human interference. The procedure of reducing the pavement video image into quantifiable distress data involves a number of steps that are difficult to quantify. The image processing algorithms adopted for the pavements always need to be calibrated using the field data for various field conditions. Currently, there is no simple and objective statistical procedure available for equitably evaluating various automated devices on the basis of resulting data. A basic statistical measurement model for estimating errors of measurement addresses, at least, the accuracy and precision of automated distress measuring devices. Factors that must be considered in any statistical decision-making process include paired versus individual measurements, threshold versus stipulated level of significance concept, and Type I versus Type II errors. Because the statistical measurement model described is general, it can be applied for evaluating any automated pavement distress measuring device.
Samenvatting