Automated monitoring of vehicle exits
The present disclosure provides techniques for machine learning-based monitoring of restricted objects. A set of images depicting an exit on an aircraft is accessed, and at least one image of the set of images is processed using a first machine learning model to determine an orientation of at least a portion of an individual relative to the exit. In response to determining that the orientation is towards the exit, it is determined whether one or more alert criteria are satisfied, where the one or more alert criteria comprise a threshold period of time. In response to determining that the one or more alert criteria are satisfied, one or more alerts indicating the exit are generated.
1 . A method, comprising:
accessing a set of images depicting an exit on an aircraft;
processing at least one image of the set of images using a first machine learning model to determine an orientation of a foot of an individual relative to the exit;
in response to determining that the orientation is towards the exit, determining whether one or more alert criteria are satisfied, wherein the one or more alert criteria comprise a threshold period of time; and
in response to determining that the one or more alert criteria are satisfied, generating one or more alerts indicating the exit.
2 . The method of claim 1 , wherein determining the orientation of the foot of the individual comprises:
identifying a toe keypoint and a heel keypoint; and
determining that the toe keypoint is relatively closer to the exit, as compared to the heel keypoint.
3 . The method of claim 1 , further comprising determining an orientation of a hand of the individual.
4 . The method of claim 3 , wherein;
determining that the orientation is towards the exit comprises determining that a back of the hand of the individual is visible in the at least one image; and
determining that the one or more alert criteria are satisfied comprises determining that the hand of the individual overlaps with a handle of the exit.
5 . The method of claim 1 , wherein generating one or more alerts comprises at least one of:
(i) activating a vibration alarm on a device of a crewmember of the aircraft;
(ii) activating a warning display on a device of a crewmember of the aircraft;
(iii) activating a vibration alarm on a seat of a passenger of the aircraft, wherein the seat of the passenger is within a defined proximity to the exit; or
(iv) activating a warning display on a seat display of the aircraft, wherein the seat display is within a defined proximity to the exit.
6 . The method of claim 1 , further comprising, in response to determining that the one or more alert criteria are no longer satisfied, terminating the one or more alerts.
7 . The method of claim 1 , further comprising:
processing at least one image of the set of images using a second machine learning model to determine a positioning of an object relative to the exit; and
in response to determining, based on the positioning, that the object obstructs at least a threshold percentage of the exit, generating one or more second alerts indicating the exit.
8 . A system comprising:
one or more memories collectively storing computer-executable instructions; and
one or more processors configured to collectively execute the computer-executable instructions and cause the system to perform an operation comprising:
accessing a set of images depicting an exit on an aircraft;
processing at least one image of the set of images using a first machine learning model to determine an orientation of a foot of an individual relative to the exit;
in response to determining that the orientation is towards the exit, determining whether one or more alert criteria are satisfied, wherein the one or more alert criteria comprise a threshold period of time; and
in response to determining that the one or more alert criteria are satisfied, generating one or more alerts indicating the exit.
9 . The system of claim 8 , wherein determining the orientation of the foot of the individual comprises:
identifying a toe keypoint and a heel keypoint; and
determining that the toe keypoint is relatively closer to the exit, as compared to the heel keypoint.
10 . The system of claim 9 , further comprising determining an orientation of a hand of the individual.
11 . The system of claim 10 , wherein;
determining that the orientation is towards the exit comprises determining that a back of the hand of the individual is visible in the at least one image; and
determining that the one or more alert criteria are satisfied comprises determining that the hand of the individual overlaps with a handle of the exit.
12 . The system of claim 8 , wherein generating one or more alerts comprises at least one of:
(i) activating a vibration alarm on a device of a crewmember of the aircraft;
(ii) activating a warning display on a device of a crewmember of the aircraft;
(iii) activating a vibration alarm on a seat of a passenger of the aircraft, wherein the seat of the passenger is within a defined proximity to the exit; or
(iv) activating a warning display on a seat display of the aircraft, wherein the seat display is within a defined proximity to the exit.
13 . The system of claim 8 , the operation further comprising:
processing at least one image of the set of images using a second machine learning model to determine a positioning of an object relative to the exit; and
in response to determining, based on the positioning, that the object obstructs at least a threshold percentage of the exit, generating one or more second alerts indicating the exit.
14 . A method, comprising:
accessing a set of images depicting an exit on an aircraft;
processing at least one image of the set of images using a first machine learning model to determine an orientation of a hand of an individual relative to the exit;
in response to determining that a back of the hand of the individual is visible in the at least one image, determining that the orientation is towards the exit;
in response to determining that the orientation is towards the exit, determining whether one or more alert criteria are satisfied;
in response to determining that the hand of the individual overlaps with a handle of the exit, determining that the one or more alert criteria are satisfied; and
in response to determining that the one or more alert criteria are satisfied, generating one or more alerts indicating the exit.
15 . The method of claim 14 , wherein the one or more alert criteria comprise a threshold period of time.
16 . The method of claim 14 , wherein generating one or more alerts comprises at least one of:
(i) activating a vibration alarm on a device of a crewmember of the aircraft;
(ii) activating a warning display on a device of a crewmember of the aircraft;
(iii) activating a vibration alarm on a seat of a passenger of the aircraft, wherein the seat of the passenger is within a defined proximity to the exit; or
(iv) activating a warning display on a seat display of the aircraft, wherein the seat display is within a defined proximity to the exit.
17 . The method of claim 14 , further comprising, in response to determining that the one or more alert criteria are no longer satisfied, terminating the one or more alerts.
18 . A method, comprising:
accessing a set of images depicting an exit on an aircraft;
processing at least one image of the set of images using a first machine learning model to determine an orientation of at least a portion of an individual relative to the exit;
in response to determining that the orientation is towards the exit, determining whether one or more alert criteria are satisfied;
in response to determining that the one or more alert criteria are satisfied, generating one or more alerts indicating the exit;
processing at least one image of the set of images using a second machine learning model to determine a positioning of an object relative to the exit; and
in response to determining, based on the positioning, that the object obstructs at least a threshold percentage of the exit, generating one or more second alerts indicating the exit.
19 . The method of claim 18 , wherein the one or more alert criteria comprise a threshold period of time.
20 . The method of claim 17 , wherein generating one or more alerts comprises at least one of:
(i) activating a vibration alarm on a device of a crewmember of the aircraft;
(ii) activating a warning display on a device of a crewmember of the aircraft;
(iii) activating a vibration alarm on a seat of a passenger of the aircraft, wherein the seat of the passenger is within a defined proximity to the exit; or
(iv) activating a warning display on a seat display of the aircraft, wherein the seat display is within a defined proximity to the exit.