Change Detection and Incremental Updates for Multi-Source Road Networks Considering Topological Consistency

Vector road networks are vital components of intelligent transportation systems and electronic navigation maps. There is a pressing need for efficient and rapid dynamic updates for road network data. In this paper, we propose a series of methods designed specifically for geometric change detection and the topological consistency updating of multi-source vector road networks without relying on complicated road network matching. For geometric change detection, we employ buffer analysis to compare various sources of vector road networks, differentiating between newly added, deleted, and unchanged road features. Furthermore, we utilize road shape similarity analysis to detect and recognize partial matching relationships between different road network sources. For incremental updates, we define topology consistency and propose three distinct methods for merging road nodes, aiming to preserve the topological integrity of the road network to the greatest extent possible. To address geometric conflicts and topological inconsistencies, we present a fusion and update method specifically tailored for partially matched road features. In order to verify the proposed methods, a road central line network with a scale of 1:10000 from the official institution is employed to geometrically update the commercial navigation road network of a similar scale in the remote area. The experiment results indicate that our method achieves an impressive 91.7% automation rate in detecting geometric changes for road features. For the remaining 8.3% of road features, our method provides suggestions on potential geometric changes, albeit necessitating manual verification and assessment. In terms of the incremental updating of the road network, approximately 89.2% of the data can be seamlessly updated automatically using our methods, while a minor 10.8% requires manual intervention for road updates. Collectively, our methods expedite the updating cycle of vector road network data and facilitate the seamless sharing and integrated utilization of multi-source road network data.