Damage prediction system using artificial intelligence
View Patent ↗A damage prediction system that uses hazard data and/or aerial images to predict future damage and/or estimate existing damage to a structure is described herein. For example, the damage prediction system may use forecasted hazard data to predict future damage or use actual hazard data to estimate existing damage. The damage prediction system may obtain hazard data in which structures were or will be impacted by a hazard. The damage prediction system can then generate a flight plan that causes an aerial vehicle to fly over the impacted parcels and capture images. The damage prediction system can use artificial intelligence to process the images for the purpose of identifying potential damage. The damage prediction system can also use a hazard model, the hazard data, and structure characteristics to generate a damage score. The damage prediction system can then use the processed images and/or damage score to generate a virtual claim.
1. A system for estimating structural damage to a physical object at a geographic location impacted by a hazard using artificial intelligence, the system comprising:
memory storing computer-executable instructions; and
a processor, wherein the computer-executable instructions, when executed by the processor, cause the processor to at least:
train an artificial intelligence model using a training set of images, wherein one or more pixels of one or more images in the training set of images is annotated to indicate structural damage to a second physical object;
obtain hazard data that indicates the hazard that impacted the geographic location;
obtain an image of the physical object at the geographic location impacted by the hazard;
apply the image of the physical object at the geographical location impacted by the hazard as an input to the trained artificial intelligence model; and
output an annotated image from the trained artificial intelligence model including an annotation of the structural damage to the physical object at the geographic location impacted by the hazard.
2. The system of claim 1 , wherein annotation of the one or more pixels comprises a modification to one attribute of the one or more pixels.
3. The system of claim 1 , wherein annotation of the one or more pixels comprises text that indicates that the one or more pixels is associated with damage to the second physical object.
4. The system of claim 3 , wherein the text further comprises an indication of a coordinate that identifies a location of the one or more pixels in the one or more images in the training set of images.
5. The system of claim 1 , wherein the trained artificial intelligence model is associated with a first type of hazard, and wherein the training set of images are each associated with the first type of hazard.
6. The system of claim 1 , wherein the image is obtained from one of an aerial vehicle or a social network.
7. The system of claim 1 , wherein the hazard data comprises at least one of hail data, wind data, flood data, storm surge data, lightning strike data, tornado data, hurricane data, earthquake data, or wildfire data.
8. A computer-implemented method of estimating structural damage to a physical object at a geographic location impacted by a hazard using artificial intelligence, the computer-implemented method comprising:
training an artificial intelligence model using a training set of images, wherein one or more pixels of one or more images in the training set of images is annotated to indicate structural damage to a second physical object;
obtaining hazard data that indicates the hazard that impacted the geographic location;
obtaining an image of the physical object at the geographic location impacted by the hazard; and
applying the image of the physical object at the geographic location impacted by the hazard as an input to the trained artificial intelligence model; and
outputting an annotated image from the trained artificial intelligence model including an annotation of the structural damage to the physical object at the geographic location impacted by the hazard.
9. The computer-implemented method of claim 8 , wherein annotation of the one or more pixels comprises a modification to one attribute of the one or more pixels.
10. The computer-implemented method of claim 8 , wherein annotation of the one or more pixels comprises text that indicates that the one or more pixels is associated with damage to the second physical object.
11. The computer-implemented method of claim 10 , wherein the text further comprises an indication of a coordinate that identifies a location of the one or more pixels in the one or more images in the training set of images.
12. The computer-implemented method of claim 8 , wherein the trained artificial intelligence model is associated with a first type of hazard, and wherein the training set of images are each associated with the first type of hazard.
13. The computer-implemented method of claim 8 , wherein the image is obtained from one of an aerial vehicle or a social network.
14. The computer-implemented method of claim 8 , wherein the hazard data comprises at least one of hail data, wind data, flood data, storm surge data, lightning strike data, tornado data, hurricane data, earthquake data, or wildfire data.
15. Non-transitory, computer-readable storage media comprising computer-executable instructions for estimating structural damage to a physical object at a geographic location impacted by a hazard using artificial intelligence, wherein the computer-executable instructions, when executed by a computer system, cause the computer system to:
train an artificial intelligence model using a training set of images, wherein one or more pixels of one or more images in the training set of images is annotated to indicate structural damage to a second physical object;
obtain hazard data that indicates the hazard that impacted the geographic location;
obtain an image of the physical object at the geographic location impacted by the hazard; and
apply the image of the physical object at the geographic location impacted by the hazard as an input to the trained artificial intelligence model; and
output an annotated image from the trained artificial intelligence model including an annotation of the structural damage to the physical object at the geographic location impacted by the hazard.
16. The non-transitory, computer-readable storage media of claim 15 , wherein annotation of the one or more pixels comprises a modification to one attribute of the one or more pixels.
17. The non-transitory, computer-readable storage media of claim 15 , wherein annotation of the one or more pixels comprises text that indicates that the one or more pixels is associated with damage to the second physical object.
18. The non-transitory, computer-readable storage media of claim 17 , wherein the text further comprises an indication of a coordinate that identifies a location of the one or more pixels in the one or more images in the training set of images.
19. The non-transitory, computer-readable storage media of claim 15 , wherein the trained artificial intelligence model is associated with a first type of hazard, and wherein the training set of images are each associated with the first type of hazard.
20. The non-transitory, computer-readable storage media of claim 15 , wherein the image is obtained from one of an aerial vehicle or a social network.