3D Map Geocoding Update via Displacement Data Registration

The patent segments the update process into two distinct parts: (1) a full recalculation phase performed periodically or initially to establish an accurate baseline 3D map representation, and (2) a incremental update phase that uses only displacement data determination for integrating new images. This segmentation allows the system to achieve both high accuracy and processing efficiency by avoiding repeated full recalculations.

The patent performs preliminary full recalculation to establish an accurate initial 3D map representation and geocoded reference data structure. This preliminary action creates a solid foundation that enables subsequent incremental updates to be performed efficiently without sacrificing accuracy, as the detailed spatial relationships and geocoding framework are already in place.

The patent introduces displacement data as an intermediary element that bridges new images and the existing 3D map representation. Instead of directly integrating new images (which would require time-consuming full recalculations), the system first computes displacement data that captures the geometric transformations, then applies these transformations to update the geocoded reference data. This intermediary approach enables fast updates while preserving accuracy.

The patent changes the approach from updating the entire 3D map representation structure to updating only the displacement parameters that describe geometric transformations. By focusing on parameter changes (displacement fields) rather than structural recalculations, the system achieves rapid integration of new images while maintaining the reliability of the underlying geocoded reference data framework.

The patent applies partial action by determining displacement data only for the specific 3D positions that are affected by new image integration, rather than recalculating for all positions in the 3D map representation. This selective approach computes displacement data for a subset of positions where changes are needed, reducing computational complexity while maintaining quality where it matters most.

The patent implements local quality by determining displacement data with high precision only in regions where new images are integrated, while other regions retain their existing geocoded reference data without requiring recalculation. This localized approach ensures high quality where needed while minimizing overall computational complexity.