Back to results
US 12,592,039 B2Granted
Vector data projection and feature matching to determine three-dimensional structure
AssigneeEcopia Tech Corporation
InventorsYuanming Shu, Shuo Tan, Zihao Chen, Ruijie Deng
Life of this patent
Estimating the term…
Abstract
Methods and systems for determining the height of structures based on imagery of the structures and associated two-dimensional vector data are provided. An example method involves projecting two-dimensional vector data outlining a roof of a structure into images of the structure captured from different perspectives and feature matching the vector data across the imagery to determine a best-matching three-dimensional position for the roof situated at the height of the structure.
Claims19
A method comprising:
- accessing georeferenced vector data that represents an outline around a roof of a building that was generated based on a reference image;
- accessing multiview imagery that depicts the building from multiple points of view;
- projecting the georeferenced vector data into the multiview imagery at a series of proposed positions within each image of at least a subset of the multiview imagery, wherein each of the proposed positions are spaced apart along a line passing through three-dimensional space;
- applying a machine learning model to generate feature descriptors corresponding to the georeferenced vector data as projected into the reference image and the georeferenced vector data as projected into the multiview imagery at each of the proposed positions;
- determining similarity scores for the feature descriptors of the georeferenced vector data at each of the proposed positions, in each of the images of the subset of multiview imagery, as compared to the feature descriptors of the georeferenced vector data in the reference image;
- and
- determining, based on the similarity scores, a best matching position for the georeferenced vector data that is most closely aligned with a true height of the outline of the roof of the building.
4 more elements
A system comprising one or more computing devices configured to:
- access georeferenced vector data that represents an outline around a roof of a building that was generated based on a reference image;
- access multiview imagery that depicts the building from multiple points of view;
- project the georeferenced vector data into the multiview imagery at a series of proposed positions within each image of at least a subset of the multiview imagery, wherein each of the proposed positions are spaced apart along a line passing through three-dimensional space;
- apply a machine learning model to generate feature descriptors corresponding to the georeferenced vector data as projected into the reference image and the georeferenced vector data as projected into the multiview imagery at each of the proposed positions;
- determine similarity scores for the feature descriptors of the georeferenced vector data at each of the proposed positions, in each of the images of the subset of multiview imagery, as compared to the feature descriptors of the georeferenced vector data in the reference image;
- and
- determine, based on the similarity scores, a best matching position for the georeferenced vector data that is most closely aligned with a true height of the outline of the roof of the building.
4 more elements
At least one non-transitory machine-readable storage medium comprising instructions that when executed cause one or more processors to:
- access georeferenced vector data that represents an outline around a roof of a building that was generated based on a reference image;
- access multiview imagery that depicts the building from multiple points of view;
- project the georeferenced vector data into the multiview imagery at a series of proposed positions within each image of at least a subset of the multiview imagery, wherein each of the proposed positions are spaced apart along a line passing through three-dimensional space;
- apply a machine learning model to generate feature descriptors corresponding to the georeferenced vector data as projected into the reference image and the georeferenced vector data as projected into the multiview imagery at each of the proposed positions;
- determine similarity scores for the feature descriptors of the georeferenced vector data at each of the proposed positions, in each of the images of the subset of multiview imagery, as compared to the feature descriptors of the georeferenced vector data in the reference image;
- and
- determine, based on the similarity scores, a best matching position for the georeferenced vector data that is most closely aligned with a true height of the outline of the roof of the building.
4 more elements
Dependencies are inferred from the claim text.