Polygon-Based Place Indexing for Accurate Proximity Detection

The patent segments a place into multiple discrete sample points distributed within its geographic boundary. Each sample point is independently identifiable and can be used for distance calculations. This segmentation allows the system to evaluate multiple locations within a place rather than relying on a single representative point, thereby improving measurement precision while maintaining manageable data complexity through structured indexing.

The patent transitions from representing a place as a zero-dimensional point to representing it as a two-dimensional area covered by multiple sample points. This dimensional expansion allows the system to capture the spatial extent of places and perform more accurate proximity assessments by checking whether user locations fall within the polygon boundaries defined by sample points.

The patent performs preliminary actions by pre-computing and storing forward indices that map places to their constituent sample points, and inverted indices that map sample points back to places. These pre-computed indices are stored in data structures that enable efficient querying. When determining nearby places, the system can directly consult these pre-prepared indices rather than performing complex geometric calculations at query time, thus improving reliability without proportionally increasing operational complexity.

The patent introduces sample points as intermediary elements between user locations and places. Instead of directly comparing user coordinates with place boundaries, the system uses sample points as mediators: it checks whether user locations are proximate to any sample points, and if so, retrieves the associated place through the inverted index. This intermediary approach simplifies the computational process while maintaining accurate place identification.