Development and Validation of Semi-Automated Software for the Analysis of Pavement Distresses

An infrastructure management company did a complete research on the implementation and validation of a digital distress data collection system in Chile, sponsored by a governmental fund. The following document presents the development and validation of a semi-automated software, developed by the study, for the analysis of pavement distress digital images. The main scope of the study is to present the experience of the software development, present the results of a precision study comparing the semi-automated analysis to manual analysis, and therefore, make recommendations on the application and implementation of pavement distress semi-automated analysis. The validation was performed using Strategic Highway Research Program (SHRP) Distress Identification Manual and a local protocol of the Ministry of Public Works of Chile. Manual inspection and image semi-automated analysis of asphalt and jointed concrete test sections were done by two evaluators. Precision of digital image technology compared to manual inspection and variability between operators using both methodologies is analyzed and discussed in this paper. Semi-automated image analysis software is a feasible and reliable alternative to traditional manual inspection applied at project and network level. Variability between evaluators was reduced when using semi-automated digital image analysis, especially on cracks and surface defects having high appearance in test sections. Asphalt pavements presented lower precision compared to jointed concrete pavement. It is more difficult to recognize distresses in dark pavements having low crack-pavement color contrast. Finally, a simple quality control/quality assurance (QC/QA) methodology is proposed to improve the application of digital technology using semi-automated image analysis.