Method and system for automatically detecting a fence
AssigneeUnited Services Automobile Association (USAA)
InventorsKelly Q. Baker, Sumita T. Jonak, Will Kerns Maney
Life of this patent
Estimating the term…
Abstract
A system and method for automatically detecting fences on a premises are disclosed. The system and method use image data gathered from one or more sources, including satellite images and images from ground-based cameras. The images are analyzed using a machine learning system to detect a fence. The system and method can also automatically update a policy holder's insurance policy to reflect the presence of a fence on the premises.
Claims20
A method of analyzing images of a premises by a computing device to determine whether a fence is present at the premises, the method comprising:
- retrieving, at the computing device, at least one image of the premises captured from a camera disposed proximally to the premises;
- comparing, by the computing device, geotag information associated with the at least one image of the premises to a known location of the premises retrieved from an insurance policy associated with the premises;
- upon confirming that the geotag information matches the known location of the premises, analyzing the at least one image of the premises using a machine learning system, wherein an output of the machine learning system includes a fence classification parameter;
- determining, by the computing device, whether or not the premises has a fence based on the fence classification parameter;
- and
- wherein, upon determining that a fence is present at the premises, modifying the insurance policy associated with the premises in response.
A method of analyzing images of a premises by a computing device to determine whether a fence is present at the premises, the method comprising:
- retrieving at least one image of the premises from a satellite images database, the at least one image corresponding to an overhead view of the premises;
- analyzing the at least one image of the premises using a machine learning system, wherein an output of the machine learning system includes a fence classification parameter;
- determining, by the computing device, whether or not the premises has a fence based on the fence classification parameter;
- retrieving at least one image of the premises captured from a camera disposed proximally to the premises;
- analyzing the at least one image of the premises captured from the camera using the machine learning system to determine whether or not the premises has a fence;
- and
- modifying the insurance policy associated with the premises to indicate a fence is present only when the presence of the fence is determined from both (1) the at least one image from the satellite images database, and (2) the at least one image captured by the camera.
Dependencies are inferred from the claim text.