Improving road network databases by integrating a geographic information system and digital imagery
Road lighting information is an important record in road network databases. It has been frequently observed that such data are either missing or not updated due to the high data collection cost using traditional methods. This paper presents a new methodology for identifying and (or) resolving missing and conflicting road lighting data in road network databases. The methodology is based on: (i) integrating the single-line road network (SLRN) in a geographic information system format with a road network database and (ii) integrating the SLRN with a collision database. Missing and conflicting lighting data are resolved using a semi-automatic method for extracting streetlight pole information. The integrated system can also identify inconsistencies related to short segments and segments with mixed illumination characteristics. Inconsistencies in the traffic volume database were also examined and the effect of data inconsistency on safety performance functions was evaluated. The proposed methodology represents an inexpensive, efficient tool for improving the quality of road network databases and associated road safety analysis.