摘要

The combination of different data sources offers possibilities to derive new required information. However, it is a complex task to match corresponding features from two datasets differing in scales and acquisition time. Taking datasets from Beijing's prime farmland survey as an example, relation of corresponding features are examined and a new relation named contain is introduced. Most current algorithms fail to identify features standing for the prime farmland efficiently for their low universality, especially in detecting corresponding features in contain relation. To improve the efficiency of feature matching, an approach is proposed in this paper which makes use of the overlay area combined with topological-difference distance with the thresholds obtained through experimentation. The approach can make the feature matching of Beijing's prime farmland survey process automatically with high accuracy and provide references for other similar matching cases.

全文