Detecting Corresponding Vertex Pairs between Planar Tessellation Datasets with Agglomerative Hierarchical Cell-Set...

Area of Science:

• Geographic Information Science
• Computer Vision
• Computational Geometry

Background:

• Accurate registration of spatial datasets is crucial for various applications, including cadastral mapping and topographical analysis.
• Existing methods for matching planar tessellation datasets often struggle with local positional discrepancies.

Purpose of the Study:

• To develop a robust method for detecting corresponding vertex pairs between planar tessellation datasets.
• To improve the accuracy of map transformation by addressing challenges posed by local positional variations.

Main Methods:

• Utilizes agglomerative hierarchical co-clustering to identify corresponding cell-set pairs.
• Detects corresponding vertex pairs from identified cell-set pairs.
• Performs map transformation based on the detected vertex pairs.

Main Results:

• The proposed method achieved an F-measure of 0.84 on complicated synthetic cadastral and topographical map datasets.
• Demonstrated significantly improved matching performance compared to a previous method which obtained an F-measure of 0.48.
• Showcased resilience to locally uneven positional discrepancies between datasets.

Conclusions:

• The developed method effectively detects corresponding vertex pairs in planar tessellation datasets.
• The approach offers superior matching performance and robustness, particularly for complex spatial data.
• This technique holds potential for enhancing accuracy in geospatial data integration and analysis.