Identifying suspicious entities using autonomous vehicles
View Patent ↗Systems and methods for identifying suspicious entities using autonomous vehicles are disclosed. In one embodiment, a method is disclosed comprising identifying a suspect vehicle using at least one digital camera equipped on an autonomous vehicle; identifying a set of candidate autonomous vehicles; enabling, on each of the candidate autonomous vehicles, a search routine, the search routine instructing each respective autonomous vehicle to coordinate tracking of the suspect vehicle; recording, while tracking the suspect vehicle, a plurality of images of the suspect vehicle; periodically re-calibrating the search routines executed by the autonomous vehicles based on the plurality of images; and re-routing the autonomous vehicles based on the re-calibrated search routines.
1 . An apparatus comprising:
a processor; and
a storage medium for tangibly storing thereon program logic for execution by the processor, the program logic comprising:
logic, executed by the processor, for receiving suspect vehicle data (SVD) associated with a suspect vehicle, wherein the SVD comprises identification data of the suspect vehicle obtained from an external source;
logic, executed by the processor, for receiving a set of images, the set of images captured by a plurality of sensors installed on a plurality of autonomous vehicles coordinating to track the suspect vehicle;
logic, executed by the processor, for inputting the set of images into a deep neural network (DNN), the DNN configured to output a set of vehicle features extracted from the images;
logic, executed by the processor, for determining that at least one feature in the set of vehicle features matches the SVD; and
logic, executed by the processor, for instructing at least one autonomous vehicle to navigate to a location of the suspect vehicle.
2 . The apparatus of claim 1 , wherein the SVD comprises one or more of a license plate number, make, model, or color of the suspect vehicle.
3 . The apparatus of claim 2 , wherein the set of vehicle features comprises one or more of a license plate number, make, model, or color of vehicles appearing in the set of images.
4 . The apparatus of claim 1 , wherein the plurality of sensors comprises a plurality of cameras.
5 . The apparatus of claim 1 , wherein the set of images comprises a set of three-dimensional images.
6 . The apparatus of claim 1 , wherein the program logic further comprises logic, executed by the processor, for receiving the SVD comprises inputting images of vehicles into a second DNN, the second DNN trained to classify an image as including a suspect vehicle and selecting an output of the second DNN as the SVD.
7 . A method comprising:
receiving, by a processor, suspect vehicle data (SVD) associated with a suspect vehicle, wherein the SVD comprises identification data of the suspect vehicle obtained from an external source;
receiving, by the processor, a set of images, the set of images captured by a plurality of sensors installed on a plurality of autonomous vehicles coordinating to track the suspect vehicle;
inputting, by the processor, the set of images into a deep neural network (DNN), the DNN configured to output a set of vehicle features extracted from the images;
determining, by the processor, that at least one feature in the set of vehicle features matches the SVD; and
instructing, by the processor, at least one autonomous vehicle to navigate to a location of the suspect vehicle.
8 . The method of claim 7 , wherein the SVD comprises one or more of a license plate number, make, model, or color of the suspect vehicle.
9 . The method of claim 8 , wherein the set of vehicle features comprises one or more of a license plate number, make, model, or color of vehicles appearing in the set of images.
10 . The method of claim 7 , wherein the plurality of sensors comprises a plurality of cameras.
11 . The method of claim 7 , wherein the set of images comprises a set of three-dimensional images.
12 . The method of claim 7 , wherein receiving the SVD comprises inputting images of vehicles into a second DNN, the second DNN trained to classify an image as including a suspect vehicle and selecting an output of the second DNN as the SVD.
13 . A non-transitory computer-readable storage medium for tangibly storing computer program instructions capable of being executed by a computer processor, the computer program instructions defining steps of:
receiving, by a processor, suspect vehicle data (SVD) associated with a suspect vehicle, wherein the SVD comprises identification data of the suspect vehicle obtained from an external source;
receiving, by the processor, a set of images, the set of images captured by a plurality of sensors installed on a plurality of autonomous vehicles coordinating to track the suspect vehicle;
inputting, by the processor, the set of images into a deep neural network (DNN), the DNN configured to output a set of vehicle features extracted from the images;
determining, by the processor, that at least one feature in the set of vehicle features matches the SVD; and
instructing, by the processor, at least one autonomous vehicle to navigate to a location of the suspect vehicle.
14 . The non-transitory computer-readable storage medium of claim 13 , wherein the SVD comprises one or more of a license plate number, make, model, or color of the suspect vehicle.
15 . The non-transitory computer-readable storage medium of claim 14 , wherein the set of vehicle features comprises one or more of a license plate number, make, model, or color of vehicles appearing in the set of images.
16 . The non-transitory computer-readable storage medium of claim 13 , wherein the plurality of sensors comprises a plurality of cameras.
17 . The non-transitory computer-readable storage medium of claim 13 , wherein the set of images comprises a set of three-dimensional images.
18 . The non-transitory computer-readable storage medium of claim 13 , wherein receiving the SVD comprises inputting images of vehicles into a second DNN, the second DNN trained to classify an image as including a suspect vehicle and selecting an output of the second DNN as the SVD.
19 . The non-transitory computer-readable storage medium of claim 13 , the steps further comprising generating a suspect vehicle identification file including the at least one feature and transmitting the suspect vehicle identification file to a central server for confirmation.