Eliminating duplicate and inconsistent information in OpenStreetMap geographic data

Abstract

Within the last few years, fast progress of information and navigation technologies gave rise to establishment and rapid growth of personal navigation systems. Whereas hardware, software and algorithm-specific issues successfully find their solutions, mapping remains to be a major challenge. And open mapping services appear to offer an effective solution. One of the most popular services is OpenStreetMap project, offering an open platform for creation of maps by users all over the world. Map developers use satellite images, personal GPS devices, aerial photography, or their personal terrain knowledge. However, weak formalization of mapping process results in considerable complications for third party users. The key problem is caused by the existence of full or partial duplicate objects, as well as conflicting data on objects. This article describes the R-tree solution to the problem. The article gives case studies for the described approach applied both for model data, and real data, which illustrate its advantages and effectiveness.