US 12,718,554 B2Granted

Detecting electrical grid assets

AssigneeX Development LLC

InventorsXin-Jing Wang, Anthony Ha, Sze Mei Cat Wong, Reuben Nahouraii, Arthur Robert Pope, Om Prakash Ravi

Life of this patent

Estimating the term…

Loading patent…

Abstract

Methods, systems, and apparatus, including computer programs encoded on a storage device, for mapping an electrical grid are disclosed. A method includes sampling multiple locations within a geographic region, executing a detection process for each location, the detection process including applying the set of images for the location as input to a machine learning (ML) model that is trained to identify electrical grid assets depicted within images taken from a combination of different perspectives and obtaining an output from the machine learning model that indicates whether a same electrical grid asset is identified in each of the images of the location. In response to an ML output that indicates a positive identification of the same electrical grid asset being depicted in a particular set of images of a particular location, the method further includes: selecting a number of sublocations within the region, and executing the detection process for each sublocation.

Claims19

Dependent claims

A method of training a machine learning model to identify electrical grid assets, the method comprising:

obtaining labeled training data comprising first sets of images, each first set of images comprising at least one image from a first perspective, at least one image from a second perspective, and at least one image from a third perspective, wherein each image in each first set of images comprises a label indicating whether an electrical grid asset is present in the image, and wherein images in each first set of images are grouped based on representing a common geographic region;
obtaining unlabeled training data comprising second sets of images, each second set of images comprising at least one image from the first perspective, at least one image from the second perspective, and at least one image from the third perspective, and wherein images in each second set of images are grouped based on representing a common geographic region, and wherein the labeled training data and the unlabeled training data each comprise a respective matrix of training images comprising multiple rows of training data for an electrical grid asset, each row comprising a different combination of images from the first, second, or third perspectives;
and
training a machine learning model to correlate electrical grid assets in a combination of images representing common geographic locations taken from different perspectives, the training comprising applying the unlabeled training data and the labeled training data as training input to the machine learning model.

1 more element

An electrical grid mapping method comprising:

sampling multiple locations within a geographic region by obtaining a set of images of each location, each set of images comprising at least one image of the location from a first perspective, at least one image of the location from a second perspective, and at least one image of the location from a third perspective;
executing a detection process for each location, the detection process comprising applying the set of images for the location as input to a machine learning (ML) model that is trained to identify electrical grid assets depicted within images taken from a combination of different perspectives and obtaining a respective output for the location from the machine learning model that indicates whether a same electrical grid asset is identified in each of the images of the location;
in response to determining that the respective output for a particular location of the multiple locations indicates a positive identification of a same respective electrical grid asset being depicted in the set of images for the particular location:
selecting a number of sublocations adjacent to the particular location, wherein each sublocation covers a different area of the geographic region than the particular location;
and
executing the detection process for each sublocation.

3 more elements

A system comprising:

at least one processor;
and a data store coupled to the at least one processor having instructions stored thereon which, when executed by the at least one processor, causes the at least one processor to perform operations comprising:
sampling multiple locations within a geographic region by obtaining a set of images of each location, each set of images comprising at least one image of the location from a first perspective, at least one image of the location from a second perspective, and at least one image of the location from a third perspective;
executing a detection process for each location, the detection process comprising applying the set of images for the location as input to a machine learning (ML) model that is trained to identify electrical grid assets depicted within images taken from a combination of different perspectives and obtaining a respective output for the location from the machine learning model that indicates whether a same electrical grid asset is identified in each of the images of the location;
in response to determining that the respective output for a particular location of the multiple locations indicates a positive identification of a same respective electrical grid asset being depicted in the set of images for the particular location:
selecting a number of sublocations adjacent to the particular location, wherein each sublocation covers a different area of the geographic region than the particular location;
and
executing the detection process for each sublocation.

5 more elements

Dependencies are inferred from the claim text.

Facts

Application: 18/226,716Filed 26 Jul 2023
Published: US 2025/0037453 A130 Jan 2025
Granted: 25 Aug 2026
Term ends: Not estimatedThe source records don’t support an estimate.

Classification9 codes

G06V 20/176
Terrestrial scenes: urban or other man-made structures
G06F 18/2155
Generating training patterns; Bootstrap methods, e.g. bagging or boosting: characterised by the incorporation of unlabelled data, e.g. multiple…
G06T 7/70
Image analysis: determining position or orientation of objects or cameras
G06V 10/7753
Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting: incorporation of unlabelled data, e.g. multiple instance…