Machine learning and computer vision solutions to seamless vehicle identification and environmental tracking therefor
A device captures a series of images over time in association with a gate, each image having a timestamp. The device determines, for a vehicle approaching the entry side, from a subset of images of the series of images featuring the vehicle, a first data set comprising a plurality of parameters that describe attributes of the vehicle by inputting the subset of images into a first machine learning model and a vehicle identifier of the vehicle by inputting images of the subset featuring a depiction of a license plate of the vehicle into a second machine learning model. The device stores the data set in association with one or more timestamps with the subset of images, determines a second data set for a second vehicle approaching the exit side, and responsive to determining that the first data set and the second data set match, instructs the gate to move.
1 . A method comprising:
capturing, using an edge device, a series of images over time in association with a moveable gate, each image having a timestamp, the moveable gate having an entry side and an exit side and blocking passage between the entry side and the exit side unless moved;
determining, by the edge device, for a vehicle approaching the entry side, from a subset of images of the series of images featuring the vehicle processed by the edge device, a first data set comprising:
a plurality of parameters that describe attributes of the vehicle by inputting the subset of images into a first machine learning model, wherein the plurality of parameters comprise identifying attributes of the vehicle and direction attributes of the vehicle; and
a vehicle identifier of the vehicle by inputting images of the subset featuring a depiction of a license plate of the vehicle into a second machine learning model;
transmitting an instruction, by the edge device, to a server remote from the edge device, for the server to store a data structure for the data set in association with one or more timestamps with the subset of images, wherein storing the data structure occurs responsive to determining that the direction attributes of the vehicle are consistent with an entry motion;
determining a second data set for a second vehicle approaching the exit side; and
responsive to a determination at the server that the first data set and the second data set match, instructing, by the edge device, the moveable gate to move.
2 . The method of claim 1 , wherein the first machine learning model is trained to output the identifying attributes of the vehicle using example data comprising images of vehicles that are labeled with one or more candidate identifying attributes.
3 . The method of claim 1 , wherein the first machine learning model is trained to output the direction attributes of the vehicle using example data comprising sequences of images of vehicles that are labeled as corresponding to an entry motion or an exit motion.
4 . The method of claim 1 , wherein the first machine learning model is trained to output the direction attributes of the vehicle using example data comprising sequences of images of vehicles that are labeled as corresponding to a given directional vector, and wherein the method further comprises determining whether the direction attributes correspond go an entry motion or an exit motion based on an output of a directional vector from the first machine learning model as compared to environmental factors surrounding the moveable gate.
5 . The method of claim 1 , wherein the vehicle identifier comprises geographical nomenclature and a string of characters.
6 . The method of claim 5 , wherein the second machine learning model is trained to identify the geographical nomenclature and the string of characters using training example images of license plates, where each of the training example images is labeled with its corresponding geographical nomenclature and string of characters.
7 . The method of claim 1 , wherein determining the vehicle identifier for first data set comprises determining that the vehicle identifier is unknown, and wherein the method further comprises:
transmitting an alert to an administrator, the alert associated with at least a portion of the subset of images, and
receiving, from the administrator, input that specifies the vehicle identifier.
8 . The method of claim 1 , further comprising:
applying computer vision to determine environmental factors around the vehicle; and
when instructing the moveable gate to move, applying parameters to the instructing based on the determined environmental factors.
9 . A non-transitory computer-readable medium with memory encoded thereon comprising instructions that, when executed by one or more processors, cause the one or more processors to perform operations, the instructions comprising instructions to:
capture, using an edge device, a series of images over time in association with a moveable gate, each image having a timestamp, the moveable gate having an entry side and an exit side and blocking passage between the entry side and the exit side unless moved;
determine, by the edge device, for a vehicle approaching the entry side, from a subset of images of the series of images featuring the vehicle processed by the edge device, a first data set comprising:
a plurality of parameters that describe attributes of the vehicle by inputting the subset of images into a first machine learning model, wherein the plurality of parameters comprise identifying attributes of the vehicle and direction attributes of the vehicle; and
a vehicle identifier of the vehicle by inputting images of the subset featuring a depiction of a license plate of the vehicle into a second machine learning model;
transmit an instruction, by the edge device, to a server remote from the edge device, for the server to store a data structure for the data set in association with one or more timestamps with the subset of images, wherein storing the data structure occurs responsive to determining that the direction attributes of the vehicle are consistent with an entry motion;
determine a second data set for a second vehicle approaching the exit side; and
responsive to a determination at the server that the first data set and the second data set match, instruct, by the edge device, the moveable gate to move.
10 . The non-transitory computer-readable medium of claim 9 , wherein the first machine learning model is trained to output the identifying attributes of the vehicle using example data comprising images of vehicles that are labeled with one or more candidate identifying attributes.
11 . The non-transitory computer-readable medium of claim 9 , wherein the first machine learning model is trained to output the direction attributes of the vehicle using example data comprising sequences of images of vehicles that are labeled as corresponding to an entry motion or an exit motion.
12 . The non-transitory computer-readable medium of claim 9 , wherein the first machine learning model is trained to output the direction attributes of the vehicle using example data comprising sequences of images of vehicles that are labeled as corresponding to a given directional vector, and wherein the instructions further comprise instructions to determine whether the direction attributes correspond go an entry motion or an exit motion based on an output of a directional vector from the first machine learning model as compared to environmental factors surrounding the moveable gate.
13 . The non-transitory computer-readable medium of claim 9 , wherein the vehicle identifier comprises geographical nomenclature and a string of characters.
14 . The non-transitory computer-readable medium of claim 13 , wherein the second machine learning model is trained to identify the geographical nomenclature and the string of characters using training example images of license plates, where each of the training example images is labeled with its corresponding geographical nomenclature and string of characters.
15 . The non-transitory computer-readable medium of claim 9 , wherein determining the vehicle identifier for first data set comprises determining that the vehicle identifier is unknown, and wherein the instructions further comprise instructions to:
transmitting an alert to an administrator, the alert associated with at least a portion of the subset of images; and
receiving, from the administrator, input that specifies the vehicle identifier.
16 . A system comprising:
memory with instructions encoded thereon; and
one or more processors that, when executing the instructions, are caused to perform operations, the operations comprising:
capturing, using an edge device, a series of images over time in association with a moveable gate, each image having a timestamp, the moveable gate having an entry side and an exit side and blocking passage between the entry side and the exit side unless moved;
determining, by the edge device, for a vehicle approaching the entry side, from a subset of images of the series of images featuring the vehicle processed by the edge device, a first data set comprising:
a plurality of parameters that describe attributes of the vehicle by inputting the subset of images into a first machine learning model, wherein the plurality of parameters comprise identifying attributes of the vehicle and direction attributes of the vehicle; and
a vehicle identifier of the vehicle by inputting images of the subset featuring a depiction of a license plate of the vehicle into a second machine learning model;
transmitting an instruction, by the edge device, to a server remote from the edge device, for the server to store a data structure for the data set in association with one or more timestamps with the subset of images, wherein storing the data structure occurs responsive to determining that the direction attributes of the vehicle are consistent with an entry motion;
determining a second data set for a second vehicle approaching the exit side; and
responsive to a determination at the server that the first data set and the second data set match, instructing, by the edge device, the moveable gate to move.