Abstract
This paper presents a probability-based multi-measure feature matching method in map conflation. Feature matching is used to determine the corresponding features in different datasets that represent analogous entities in the real world. In the proposed method, a total matching probability is computed by the weighted average of multiple measures, including positional measure, shape measure, directional measure and topological measure. The matching strategies for point features, linear features and areal features are also provided. The proposed method is implemented in a prototype for matching features from two different data sources, and is compared with traditional methods. The results demonstrate not only the practicability of using the proposed method to resolve feature matching issues in map conflation, but also its advantages compared with traditional methods in terms of matching effects.
Acknowledgements
We appreciate the anonymous reviewers very much for the constructive comments and for clarifying this manuscript. We thank Prof. Michael Goodchild for his valuable comments that helped strengthen the paper, and Dr Clay William for his patient checking of the English of the paper. This study was substantially supported by the National Natural Science Foundation of China (Project No. 40771174), Program for New Century Excellent Talents in Universities (Project No. NCET-06-0381), Foundation of Shanghai Dawn Scholarship and Rising-star Program (Project No. 07SG24 and 08QH14022), and grants from the Doctoral Program of Higher Education of China (Project No. 20070247046).