Back to results
US 11,769,387 B2Granted
Method and apparatus for detecting drowning
AssigneeLynxight Ltd.
InventorsOmer Bar-Ilan, Roee Mor
Life of this patent
Estimating the term…
Abstract
A method, system and computer program product, the method comprising: obtaining a multiplicity of images taken by one or more overhead capture devices overlooking a body of water having a surface; detecting a person depicted in at least some images from the multiplicity of images; determining for images in the images, whether the image depicts the person in substantially vertical position in the at least some images said determining performed upon one or more characteristics extracted from each image; and subject to determining that the person is not in substantially vertical position in the at least some images, taking an action.
Claims19
A method, comprising:
- obtaining a multiplicity of images taken by two or more overhead cameras overlooking a body of water having a surface, wherein the two overhead cameras comprise at least a first camera and a second camera, each of which has a captured portion of the body of water, the first camera and the second camera have at least a partial overlap between their respective captured portions of the body of water, wherein at least one of the first camera and the second camera is positioned at a height above 1.80 meters;
- detecting a person depicted in at least some images from the multiplicity of images;
- determining whether the at least some images depict the person in a substantially vertical position, wherein determining whether the person is in the substantially vertical position comprises determining whether a direction of the person as depicted in at least a subset of images coincides with a vertical direction at a location of the person as captured by the two or more overhead cameras, said determining performed upon at least one characteristic extracted from each image of the at least some images, wherein the at least one characteristic is selected from the group consisting of:
- a color histogram;
- an edge;
- an edge histogram;
- a gradient histogram;
- a contour edge between colors;
- a contour edge between shapes;
- a distinct feature;
- a unique pattern;
- a corner;
- a color co-occurrence matrix;
- and a texture, wherein said determining whether the person is in the substantially vertical position comprises:
- determining a length of a projection of the person as depicted in the at least some images, wherein the length of the projection is determined using a calibration of the first camera in which a vector between two points on an image is translated into an orientation on water surface;
- determining a height of the person as depicted in the at least some images;
- and
- subject to a ratio between the length and the height being below a predetermined value, determining that the person is in the substantially vertical position;
- and
- subject to determining that the person is not in the substantially vertical position in at least a predetermined number of images from the at least some images, performing an action, wherein said performing the action comprises sending an alert to a mobile device of a lifeguard.
17 more elements
A system comprising:
- a plurality of overhead cameras overlooking a body of water having a surface, wherein the plurality of overhead cameras comprises at least a first camera and a second camera, each of which has a captured portion of the body of water, the first camera and the second camera have at least a partial overlap between their respective captured portions of the body of water, wherein at least one of the first camera and the second camera is positioned at a height above 1.80 meters;
- a mobile device associated with a lifeguard observing the body of water, the mobile device being configured to notify the lifeguard regarding a potential drowning event;
- a computing platform that includes a processor, the computing platform being configured to:
- obtain a multiplicity of images taken by the plurality of overhead cameras;
- detecting a person depicted in at least some images from the multiplicity of images;
- determining whether the at least some images depict the person in a substantially vertical position, said determining performed upon at least one characteristic extracted from each image, wherein the at least one characteristic is selected from the group consisting of:
- a color histogram;
- an edge;
- an edge histogram;
- a gradient histogram;
- a contour edge between colors;
- a contour edge between shapes;
- a distinct feature;
- a unique pattern;
- a corner;
- a color co-occurrence matrix;
- and a texture, wherein determining whether the person is in the substantially vertical position comprises:
- determining that the person is in the substantially vertical position if a direction of the person is substantially vertical in the at least some images taken by the first camera, and if the direction of the person as depicted in at least a subset of the images taken by the second camera coincides with a vertical direction at a location of the person as seen by the second camera in the subset, wherein said determining whether the person is in the substantially vertical position further comprises:
- determining a length of a projection of the person as depicted in the at least some images, wherein the length of the projection is determined using a calibration of the first camera in which a vector between two points on an image is translated into an orientation on water surface;
- determining a height of the person as depicted in the at least some images;
- and
- subject to a ratio between the length and the height being below a predetermined value, determining that the person is in the substantially vertical position and
- subject to determining that the person is not in the substantially vertical position in at least a predetermined number of images from the at least some images, performing an action, wherein said performing the action comprises sending an alert to the mobile device, thereby generating a vibrating alert notifying the lifeguard of the potential drowning event.
20 more elements
Dependencies are inferred from the claim text.