Precision Test of Cracking Surveys with the Automated Distress Analyzer

In the last few years, there have been rapid developments and advances of technologies for pavement cracking surveys for both data acquisition and data interpretation. Although the high-speed and objective means of automated data collection and processing for pavement distress survey are appealing, many agencies are reluctant to adopt the new technology because either data quality is perceived as compromised or that there is lack of experience. However, there has been a slow but growing trend that more agencies began to realize that as long as a proper balance between data quality and full automation is made, the level of error in automation is acceptable. In addition, there is a lack of commonly used cracking protocols and limited analysis of data quality. This paper attempts to design an analysis method and apply real-world data sets under varying conditions for fully automated analysis. A statistical approach is used to compare results among manual, semiautomated, and fully automated procedures. The paper also presents the automated cracking survey methodologies and protocols used in the Automated Distress Analyzer (ADA) and the Digital Highway Data Vehicle (DHDV). The quality of automated cracking data is compared with manual processing results. The variability and the quality assurance procedures in the cracking surveys are discussed. It is demonstrated that fully automated processing of pavement cracking can be as useful as manual processing in addition to providing substantial savings.