US 10,593,108 B2Granted

Converting digital aerial images into a three-dimensional representation utilizing processing clusters

AssigneeSkycatch, Inc.

InventorsManlio Francisco Barajas Hernandez, David Chen, Pablo Arturo Martinez Gonzalez, Christian Sanz

Life of this patent

Estimating the term…

Loading patent…

Abstract

Systems and methods are disclosed for more efficiently and quickly utilizing digital aerial images to generate models of a site. In particular, in one or more embodiments, the disclosed systems and methods capture a plurality of digital aerial images of a site. Moreover, the disclosed systems and methods can cluster the plurality of digital aerial images based on a variety of factors, such as visual contents, capture position, or capture time of the digital aerial images. Moreover, the disclosed systems and methods can analyze the clusters independently (i.e., in parallel) to generate cluster models. Further, the disclosed systems and methods can merge the cluster models to generate a model of the site.

Claims20

Dependent claims

A system comprising:

at least one processor;
and
at least one non-transitory computer readable storage medium storing instructions thereon that, when executed by the at least one processor, cause the system to:
identify digital aerial images of a site captured by at least one unmanned aerial vehicle;
perform, at a client device, at least one of feature detection, feature matching, three-dimensional representation generation, or georeferencing;
upon performing, at the client device at least one of the feature detection, the feature matching, the three-dimensional representation generation, or the georeferencing, perform parallel processing of the digital aerial images of the site by:
determining a measure of similarity between the digital aerial images;
determining parallel processing clusters of the digital aerial images based on the measure of similarity between the digital aerial images;
and
sending the parallel processing clusters of the digital aerial images to a plurality of computing devices, that do not include the client device, for:
performing, in parallel across the plurality of computing devices, at least one of cluster feature matching, cluster three-dimensional representation generation, cluster georeferencing, cluster three-dimensional representation refinement, cluster densification, cluster three-dimensional representation reconstruction, or cluster rendering in relation to the parallel processing clusters;
and
generating cluster models corresponding to the parallel processing clusters of the digital aerial images in parallel across the plurality of computing devices, wherein the cluster models comprise a plurality of points reflecting the site;
and
generate a model of the site by merging the cluster models generated in parallel across the plurality of computing devices using the plurality of points.

12 more elements

In a digital medium environment for generating site models from aerial images, a method for parallel processing of a model generation pipeline comprising:

receiving digital aerial images of a site captured by at least one unmanned aerial vehicle;
performing, at a client device, at least one of feature detection, feature matching, three-dimensional representation generation, or georeferencing;
upon performing, at the client device at least one of the feature detection, the feature matching, the three-dimensional representation generation, or the georeferencing, performing parallel processing of the digital aerial images of the site by:
determining parallel processing clusters of the digital aerial images by analyzing, for each digital aerial image, a capture position of the digital aerial image;
sending the parallel processing clusters of the digital aerial images to a plurality of computing devices for:
performing, in parallel across the plurality of computing devices, at least one of cluster feature matching, cluster three-dimensional representation generation, cluster georeferencing, cluster three-dimensional representation refinement, cluster densification, cluster three-dimensional representation reconstruction, or cluster rendering in relation to the parallel processing clusters;
and
generating cluster models corresponding to the clusters of the digital aerial images in parallel across the plurality of computing devices;
in response to receiving the cluster models from the plurality of computing devices, merging the cluster models;
and
generating a model of the site based on the merged cluster models.

8 more elements

A system comprising:

at least one processor;
and
at least one non-transitory computer readable storage medium storing instructions thereon that, when executed by the at least one processor, cause the system to:
identify digital aerial images of a site captured by at least one unmanned aerial vehicle;
perform, at a client device, at least one of feature detection, feature matching, three-dimensional representation generation, or georeferencing;
perform parallel processing of the digital aerial images of the site by:
determining parallel processing clusters of the digital aerial images by analyzing, for each digital aerial image, visual content portrayed in the digital aerial image;
performing, in parallel across a plurality of computing devices that do not include the client device, at least one of cluster feature matching, cluster three-dimensional representation generation, cluster georeferencing, cluster three-dimensional representation refinement, cluster densification, cluster three-dimensional representation reconstruction, or cluster rendering in relation to the parallel processing clusters;
and
generating in parallel, utilizing the plurality of computing devices, cluster models corresponding to the parallel processing clusters of the digital aerial images;
and
generate a model of the site by merging the cluster models.

9 more elements

Dependencies are inferred from the claim text.