Machine learning solutions for enforcement of sensor-detected vehicle facility infractions
A device detects, using input from one or more sensors installed at a parking facility of a plurality of parking facilities, an infraction caused by a vehicle. Responsive to detecting the infraction, the device generates a vehicle fingerprint by inputting a depiction of the vehicle into a supervised machine learning model, the depiction derived from one or more images of the vehicle captured at the parking facility, and receiving a feature vector of the vehicle as output from the supervised machine learning model, the feature vector comprising a plurality of embeddings each describing a dimension of the vehicle. The device monitors for entry of the vehicle at each of the plurality of parking facilities using the vehicle fingerprint, and, responsive to detecting entry of the vehicle at a given one of the plurality of parking facilities, triggers a remediation action.
1 . A method comprising:
detecting, using input from one or more sensors installed at a parking facility of a plurality of parking facilities, an indication of a possible infraction caused by a vehicle;
responsive to detecting the indication of the possible infraction, transmitting instructions to a moveable camera system, causing the moveable camera system to automatically navigate the movable camera system to a location of the infraction and capture an image of the vehicle;
confirming, based on the image of the vehicle captured by the moveable camera system after navigating to the location of the infraction, that an infraction has in fact occurred;
responsive to confirming that the infraction has in fact occurred, generating a vehicle fingerprint of the vehicle as a reference for additional sensors deployed in the plurality of parking facilities to automatically identify the vehicle in other locations by:
inputting a depiction of the vehicle into a supervised machine learning model, the depiction derived from one or more images of the vehicle captured at the parking facility; and
receiving a feature vector of the vehicle as output from the supervised machine learning model, the feature vector comprising a plurality of embeddings each describing a dimension of the vehicle;
based on confirming that the infraction has in fact occurred, automatically monitoring, using the additional sensors, for entry of the vehicle at each of the plurality of parking facilities using the vehicle fingerprint; and
responsive to detecting entry of the vehicle at a given one of the plurality of parking facilities, triggering a remediation action.
2 . The method of claim 1 , wherein the input comprises an indication that two or more adjacent parking spaces in the parking facility have transitioned from a vacant state to an occupied state within a threshold amount of time of one another.
3 . The method of claim 2 , wherein confirming that the infraction has in fact occurred comprises:
commanding the moveable camera system to navigate to a vantage point comprising the two or more adjacent parking spaces and capture one or more images of the two or more adjacent parking spaces;
determining whether the vehicle is occupying the two or more adjacent parking spaces; and
confirming that the infraction has in fact occurred in response to determining that the vehicle is occupying the two or more adjacent parking spaces.
4 . The method of claim 1 , wherein generating the vehicle fingerprint is performed further responsive to detecting that a license plate of the vehicle is not recognized.
5 . The method of claim 4 , wherein monitoring for entry of the vehicle comprises monitoring for the license plate of the vehicle where the license plate of the vehicle is recognized.
6 . The method of claim 1 , wherein inputting the depiction of the vehicle into the supervised machine learning model comprises:
isolating, from the one or more images, a first image portion containing the vehicle; and
excluding, from the one or more images, a second image portion that does not contain the vehicle.
7 . The method of claim 1 , wherein triggering the remediation action comprises:
determining an infraction type of the infraction; and
transmitting a remediation command resulting in the remediation action based on the infraction type.
8 . The method of claim 7 , wherein the remediation command comprises a command to raise a blocking device preventing movement of the vehicle within the parking facility.
9 . The method of claim 7 , wherein the remediation command comprises a command to initiate a communication session with a law enforcement entity.
10 . The method of claim 1 , further comprising:
determining whether the vehicle is in a candidate set of known vehicles; and
responsive to determining that the vehicle is not in the candidate set of known vehicles, generating a vehicle fingerprint corresponding to the vehicle.
11 . A non-transitory computer-readable medium comprising memory with instructions encoded thereon, the instructions comprising instructions to:
detect, using input from one or more sensors installed at a parking facility of a plurality of parking facilities, an indication of a possible infraction caused by a vehicle;
responsive to detecting the indication of the possible infraction, transmitting instructions to a movable camera system, causing the moveable camera system to automatically navigate the moveable camera system to a location of the infraction and capture an image of the vehicle;
confirm, based on the image of the vehicle captured by the moveable camera system after navigating to the location of the infraction, that an infraction has in fact occurred;
responsive to confirming that the infraction has in fact occurred, generate a vehicle fingerprint of the vehicle as a reference for additional sensors deployed in the plurality of parking facilities to automatically identify the vehicle in other locations by:
inputting a depiction of the vehicle into a supervised machine learning model, the depiction derived from one or more images of the vehicle captured at the parking facility; and
receiving a feature vector of the vehicle as output from the supervised machine learning model, the feature vector comprising a plurality of embeddings each describing a dimension of the vehicle;
based on confirming that the infraction has in fact occurred, automatically monitor, using the additional sensors, for entry of the vehicle at each of the plurality of parking facilities using the vehicle fingerprint; and
responsive to detecting entry of the vehicle at a given one of the plurality of parking facilities, trigger a remediation action.
12 . The non-transitory computer-readable medium of claim 11 , wherein the input comprises an indication that two or more adjacent parking spaces in the parking facility have transitioned from a vacant state to an occupied state within a threshold amount of time of one another.
13 . The non-transitory computer-readable medium of claim 12 , wherein the instructions to confirm that the infraction has in fact occurred comprise instructions to:
command the moveable camera system to navigate to a vantage point comprising the two or more adjacent parking spaces and capture one or more images of the two or more adjacent parking spaces;
determine whether the vehicle is occupying the two or more adjacent parking spaces; and
confirm that the infraction has in fact occurred in response to determining that the vehicle is occupying the two or more adjacent parking spaces.
14 . The non-transitory computer-readable medium of claim 11 , wherein generating the vehicle fingerprint is performed further responsive to detecting that a license plate of the vehicle is not recognized.
15 . The non-transitory computer-readable medium of claim 14 , wherein monitoring for entry of the vehicle comprises monitoring for the license plate of the vehicle where the license plate of the vehicle is recognized.
16 . The non-transitory computer-readable medium of claim 11 , wherein inputting the depiction of the vehicle into the supervised machine learning model comprises:
isolating, from the one or more images, a first image portion containing the vehicle; and
excluding, from the one or more images, a second image portion that does not contain the vehicle.
17 . The non-transitory computer-readable medium of claim 11 , wherein the instructions to trigger the remediation action comprise instructions to:
determine an infraction type of the infraction; and
transmit a remediation command resulting in the remediation action based on the infraction type.
18 . The non-transitory computer-readable medium of claim 17 , wherein the remediation command comprises a command to raise a blocking device preventing movement of the vehicle within the parking facility.
19 . The non-transitory computer-readable medium of claim 17 , wherein the remediation command comprises a command to initiate a communication session with a law enforcement entity.
20 . A system comprising:
memory with instructions encoded thereon; and
one or more processors that, when executing the instructions, are caused to perform operations comprising:
detecting, using input from one or more sensors installed at a parking facility of a plurality of parking facilities, an indication of a possible infraction caused by a vehicle;
responsive to detecting the indication of the possible infraction, transmitting instructions to a moveable camera system, causing the moveable camera system to automatically navigate the movable camera system to a location of the infraction and capture an image of the vehicle;
confirming, based on the image of the vehicle captured by the moveable camera system after navigating to the location of the infraction, that an infraction has in fact occurred;
responsive to confirming that the infraction has in fact occurred, generating a vehicle fingerprint of the vehicle as a reference for additional sensors deployed in the plurality of parking facilities to automatically identify the vehicle in other locations by:
inputting a depiction of the vehicle into a supervised machine learning model, the depiction derived from one or more images of the vehicle captured at the parking facility; and receiving a feature vector of the vehicle as output from the supervised machine learning model, the feature vector comprising a plurality of embeddings each describing a dimension of the vehicle;
based on confirming that the infraction has in fact occurred, automatically monitoring, using the additional sensors, for entry of the vehicle at each of the plurality of parking facilities using the vehicle fingerprint; and
responsive to detecting entry of the vehicle at a given one of the plurality of parking facilities, triggering a remediation action.