US 11,783,463 B2
Systems and methods for artificial intelligence (AI) roof deterioration analysis
Brad E. Gawron, Somers, CT (US); and James A. Quadrino, Glastonbury, CT (US)
Assigned to The Travelers Indemnity Company, Hartford, CT (US)
Filed by The Travelers Indemnity Company, Hartford, CT (US)
Filed on Jan. 5, 2023, as Appl. No. 18/93,770.
Application 18/093,770 is a continuation of application No. 17/136,984, filed on Dec. 29, 2020, granted, now 11,574,394.
Claims priority of provisional application 63/112,884, filed on Nov. 12, 2020.
Prior Publication US 2023/0144043 A1, May 11, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06K 9/00 (2022.01); G06T 7/00 (2017.01); G06T 7/11 (2017.01); G06N 3/02 (2006.01); G06V 10/40 (2022.01); G06F 18/2113 (2023.01); G06F 18/2431 (2023.01)
CPC G06T 7/0002 (2013.01) [G06F 18/2113 (2023.01); G06F 18/2431 (2023.01); G06N 3/02 (2013.01); G06T 7/11 (2017.01); G06T 7/97 (2017.01); G06V 10/40 (2022.01); G06T 2207/10032 (2013.01); G06T 2207/20132 (2013.01); G06T 2207/30184 (2013.01)] 17 Claims
OG exemplary drawing
 
1. An Artificial Intelligence (AI) roof deterioration analysis method, comprising:
receiving, by an electronic processing device, a high-resolution image of a property upon which a structure comprising a roof is situated, the high-resolution image being taken at a first time;
identifying, by the electronic processing device and based on a location of the property, a shape file descriptive of a shape of the roof;
trimming, by the electronic processing device and by overlaying the high-resolution image of the property and the shape of the roof, the high-resolution image, thereby defining a high-resolution roof image;
segmenting, by the electronic processing device, the high-resolution roof image into a plurality of image tiles;
identifying, by the electronic processing device and utilizing defect detection rules defined by AI roof deterioration analysis instructions stored in a non-transitory computer readable memory device in communication with the electronic processing device, and for each image tile of the plurality of image tiles, a plurality of features of the image tile;
scoring, by the electronic processing device and based on a feature scoring model of the AI roof deterioration analysis instructions, each of the plurality of features of each image tile;
computing, by the electronic processing device and utilizing scores for each of the plurality of features of each image tile, an overall score for each image tile;
computing, by the electronic processing device and utilizing the overall scores for each of the image tiles of the plurality of image tiles, a first overall defectiveness score for the roof;
outputting, by the electronic processing device, the first overall defectiveness score for the roof;
identifying, by the electronic processing device, a plurality of previous overall defectiveness scores for the roof that are based on a respective plurality of previous high-resolution images of the property taken at a plurality of previous times before the first time;
computing, by the electronic processing device and based on a comparison of the plurality of previous overall defectiveness scores to the first overall defectiveness score, a measure of deviation of the first overall defectiveness score with respect to the plurality of previous overall defectiveness scores;
comparing, by the electronic processing device, the measure of deviation to a stored deviation threshold; and
classifying, by the electronic processing device and based on the comparing, the roof as being either (i) deteriorating, in the case that the measure of deviation exceeds or equals the deviation threshold or (ii) improving, in the case that the measure of deviation falls below the deviation threshold.