A STEREOMETRIC KNOWLEDGE-BASED SYSTEM FOR MAINTENANCE OF STREET NETWORKS

The main objective of this work was to investigate the potential of in tegrating a stereometric vision system, i.e., using digital stereo ima ges, and a knowledge-based system for flexible pavement distress class ification. Classification process includes distress type, severity lev el, and options for repair. A hybrid stereo vision and knowledge-based system (called K-PAVER) was developed. The system extracts distress m easurements using a PC-based stereo vision system. Geometric surface m easurements such as point locations, distances, areas, volumes, and su rface areas could also be computed. The knowledge-based system develop ed utilizes a set of if...then rules from the PAVER system (a pavement maintenance management system for roads and streets) and related lite ratures. New parameters, including shape parameters, orientation, and some geometrical measurements, were introduced to the system in order to facilitate the distress classification process. A criterion for mai ntenance priorities based on four parameters was developed. These para meters are pavement condition index, average daily traffic, location o f distressed pavement, and street class. Surface measurements and auto matic classification decision-making were validated and tested for all distress types. The developed system gives accurate results in both t he measurement mode and the decision-making phase. This result opens t he door for a fully automated distress classification process without any human intervention.