Back to results
US 8,942,535 B1Granted
Implicit video location augmentation
AssigneeGoogle Inc.
InventorFrancis William Galligan
Life of this patent
Estimating the term…
Abstract
A method and apparatus for performing implicit video location augmentation are provided. Implicit video location augmentation may include identifying a first geolocation for a first frame from a plurality of video frames based on a first image captured by the first frame, identifying a second geolocation for a second frame from the plurality of video frames based on a second image captured by the second frame, determining, by a processor, a third geolocation for a third frame from the plurality of video frames based on the first geolocation and the second geolocation, and storing an updated plurality of video frames such that the first frame is associated with the first geolocation, the second frame is associated with the second geolocation, and the third frame is associated with the third geolocation.
Claims17
A method comprising:
- receiving an input video stream captured by an image capturing device, the input video stream including a plurality of video frames;
- identifying a first plurality of geolocations for a first frame from the plurality of video frames based on content of the first frame, wherein identifying the first plurality of geolocations includes identifying a first geolocation by:
- detecting a geolocatable feature in the content of the first frame, and
- identifying a previously geolocated orientation feature corresponding to the geolocatable feature based on a similarity between the geolocatable feature and the orientation feature, wherein the orientation feature is associated with previously identified geolocation information;
- identifying a second geolocation for a second frame from the plurality of video frames based on content of the second frame;
- determining, by a processor, a third geolocation for a third frame from the plurality of video frames based on the first geolocation and the second geolocation;
- and
- storing an updated plurality of video frames such that the first frame is associated with the first geolocation, the second frame is associated with the second geolocation, and the third frame is associated with the third geolocation.
5 more elements
A method comprising:
- receiving, from an external device, an input video captured by an external image capturing device, the input video including a plurality of video frames;
- identifying a plurality of geolocations for a first frame from the plurality of video frames based on content of the first frame by:
- detecting a geolocatable feature in the content of the first frame,
- identifying a plurality of previously geolocated orientation features corresponding to the geolocatable feature based on a similarity between the geolocatable feature and each respective orientation feature from the plurality of orientation features, wherein the respective orientation features are associated with previously identified geolocation information, and
- for each orientation feature in the plurality of orientation features, generating a confidence indicator indicating a likelihood that the respective geolocation information is accurate;
- identifying a geolocation for a second frame from the plurality of video frames based on content of the second frame;
- determining, by a processor, a geolocation for the first frame based on the plurality of geolocations identified for the first frame and the second geolocation;
- storing an updated plurality of video frames such that the first frame is associated with the first geolocation, the second frame is associated with the second geolocation, and the third frame is associated with the third geolocation;
- and
- transmitting an output video including the updated plurality of video frames to the external device.
7 more elements
A method comprising:
- receiving an input video stream captured by an image capturing device, the input video stream including a plurality of video frames;
- identifying a first geolocation for a first frame from the plurality of video frames based on content of the first frame by:
- detecting a geolocatable feature in the content of the first frame, and
- identifying a previously geolocated orientation feature corresponding to the geolocatable feature based on a similarity between the geolocatable feature and the orientation feature, wherein the orientation feature is associated with previously identified geolocation information;
- identifying a second geolocation for a second frame from the plurality of video frames based on content of the second frame by:
- detecting a second geolocatable feature in the content of the second frame, and
- identifying a second previously geolocated orientation feature corresponding to the second geolocatable feature based on a similarity between the second geolocatable feature and the second orientation feature, wherein the second orientation feature is associated with previously identified geolocation information;
- determining, by a processor, a third geolocation for a third frame from the plurality of video frames based on the first geolocation and the second geolocation;
- and
- storing an updated plurality of video frames such that the first frame is associated with the first geolocation, the second frame is associated with the second geolocation, and the third frame is associated with the third geolocation.
7 more elements
Dependencies are inferred from the claim text.