US 11,769,245 B2
Systems and methods of monitoring cargo load systems for damage detection
Nitin Kumar Goyal, Bangalore (IN); Mahesh Ainapure, Bengaluru (IN); and Ashutosh Kumar Jha, Bangalore (IN)
Assigned to GOODRICH CORPORATION, Charlotte, NC (US)
Filed by GOODRICH CORPORATION, Charlotte, NC (US)
Filed on Feb. 18, 2022, as Appl. No. 17/675,280.
Claims priority of application No. 202141047974 (IN), filed on Oct. 21, 2021.
Prior Publication US 2023/0126817 A1, Apr. 27, 2023
Int. Cl. G06T 7/00 (2017.01); H04N 7/18 (2006.01); B64F 5/60 (2017.01); G06V 10/764 (2022.01); B64D 9/00 (2006.01); G06V 10/82 (2022.01); H04N 23/66 (2023.01); H04N 23/69 (2023.01)
CPC G06T 7/0004 (2013.01) [B64D 9/00 (2013.01); B64F 5/60 (2017.01); G06V 10/764 (2022.01); G06V 10/82 (2022.01); H04N 7/183 (2013.01); H04N 23/66 (2023.01); H04N 23/69 (2023.01); G06T 2207/20084 (2013.01); G06T 2207/30108 (2013.01); G06T 2207/30232 (2013.01)] 11 Claims
OG exemplary drawing
 
1. A component inspection system for monitoring and detecting damage to components of a cargo handling system, the component inspection system comprising:
a first camera configured to monitor a first detection zone; and
an inspection system controller configured to analyze image data output by the first camera, wherein the inspection system controller is configured to:
identify a component of the cargo handling system in the image data received from the first camera, wherein the inspection system controller is configured to identify the component of the cargo handling system in the image data by:
inputting the image data into a trained component identification model; and
receiving a component identification output by the trained component identification model;
determine a state of the component, the state of the component comprising at least one of a normal state or a damaged state, wherein the inspection system controller is configured to determine the state of the component by:
selecting a trained damage classification model based on the component identification output by the trained component identification model;
inputting the image data into the trained damage classification model; and
receiving a damage classification output by the trained damage classification model, the damage classification corresponding to the state of the component;
determine a confidence score of the damage classification;
comparing the confidence score to a threshold confidence; and
command the first camera to at least one of rotate, zoom-in, or zoom-out in response to determining the confidence score to is less than the threshold confidence.