Abstract

This paper proposes an approach for automatically identifying and georeferencing street maps found on the web. The authors address the challenge of distinguishing street maps from various other types of images by employing image processing techniques, specifically Law's texture classification algorithm, to detect unique patterns such as road lines. By integrating these techniques with their past work on extracting road intersections and utilizing the GEOPPM algorithm, the approach achieves a precision of 95.45% in identifying street maps and a georeferencing precision of 71.43%, thereby demonstrating its potential utility in applications requiring accurate map search and integration.

Date

2005

Authors

Ching-Chien Chen, Sneha S Desai, Yao‐Yi Chiang, Kandarp Desai, Craig A Knoblock

Journal

Proceedings of the 2005 workshop on Geographic information retrieval