Vehicle identification system

Abstract

To provide a vehicle identification system that can prevent the occurrence of a situation in which efficient traffic is inhibited in identification of a target vehicle.SOLUTION: A vehicle identification system includes one or more processors. The one or more processors are configured to execute the following first processing to third processing. The first processing is to instruct a first vehicle that execution of a predetermined action to identify a target vehicle is started while the first vehicle travels in a vehicle identification zone. The second processing is to start recognition processing of recognizing an action executed by the first vehicle by using a sensor while the first vehicle travels in the vehicle identification zone. The third processing is to identify the first vehicle executing the predetermined action as the target vehicle.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present disclosure relates to a technique for identifying a target vehicle. [Background technology]

[0002] Identifying the target vehicle is important for providing control and services to the target vehicle. For example, in Auto Valet Parking (AVP) in a parking lot, it is required to identify the target vehicle of AVP from the viewpoint of reliability and safety.

[0003] Patent Document 1 discloses a vehicle authentication method in a parking lot as a technology for identifying a target vehicle. The vehicle authentication method disclosed in Patent Document 1 transmits information instructing the target vehicle to perform a predetermined action, checks whether the vehicle in the parking lot has performed the predetermined action, and authenticates the vehicle that has performed the predetermined action as the target vehicle and performs localization.

[0004] Other documents that demonstrate the technical level of this technical field include the following Patent Documents 2 and 3. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] U.S. Pat. No. 10,532,771 [Patent Document 2] Patent Publication No. 2021-099601 [Patent Document 3] JP 2018-112981 A Summary of the Invention [Problem to be solved by the invention]

[0006] Patent Document 1 discloses a technology for performing vehicle identification (authentication) based on the actions of the vehicle. In particular, in Patent Document 1, vehicle identification is performed with the vehicle stopped at the drop-off position. When performing vehicle identification based on the actions of the vehicle, the vehicle is required to perform an action having a certain duration, so vehicle identification takes a certain amount of time. For this reason, in the technology disclosed in Patent Document 1, the vehicle continues to be stopped at the drop-off position for a while from when vehicle identification starts until it is completed. It is undesirable for a vehicle to occupy a certain space for a long period of time, as this hinders efficient traffic.

[0007] One object of the present disclosure is to address the above-mentioned problems and to prevent the occurrence of situations in which efficient traffic is impeded in relation to technology for identifying target vehicles. [Means for solving the problem]

[0008] One aspect of the present disclosure relates to a vehicle identification system for identifying a target vehicle. The vehicle identification system includes one or more processors. The one or more processors are configured to instruct the first vehicle to start performing a predetermined action for identifying a target vehicle while the first vehicle is traveling in a vehicle identification zone, to start a recognition process for recognizing the action performed by the first vehicle using a sensor while the first vehicle is traveling in the vehicle identification zone, and to identify the first vehicle performing the predetermined action as a target vehicle. Effect of the Invention

[0009] According to the present disclosure, while the first vehicle is traveling in the vehicle identification zone, an instruction to start execution of a predetermined action is given, and the recognition process is started. As a result, when the first vehicle stops to receive the provision of the function, the target vehicle can be in a state where the identification of the target vehicle is completed, or a part of the action to be performed by the first vehicle has been recognized. Therefore, the time that the target vehicle stops for vehicle identification can be shortened. In turn, the occurrence of a situation in which efficient traffic is hindered can be suppressed. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a conceptual diagram for explaining automatic valet parking in a parking lot. [Diagram 2] 1 is a conceptual diagram for explaining an overview of a vehicle identification system according to an embodiment of the present invention. [Diagram 3] FIG. 13 is a conceptual diagram illustrating an example of a problem regarding selection of a camera to be used in recognition processing. [Figure 4] FIG. 11 is a conceptual diagram for explaining an overview of camera recognition processing. [Diagram 5] 4 is a flowchart illustrating an example of a process executed in the vehicle identification system. [Figure 6] FIG. 2 is a diagram illustrating an example of a hardware configuration of the vehicle identification system according to the present embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] 1 Introduction Various functions that provide control and services to target vehicles are being considered. One such function is Auto Valet Parking (AVP) in parking lots. AVP automatically allows target vehicles to enter and leave parking lots without user operation. Other functions include a transportation device that uses a transportation device to carry target vehicles in and out of parking lots and organize them.

[0012] FIG. 1 is a conceptual diagram for explaining AVP. The management system 100 manages the AVP in a parking lot 2. The parking lot 2 is provided with a sensor 150 that detects the position and state of each vehicle 1 in the parking lot 2. The management system 100 manages information on the parking lot 2 and each vehicle 1 based on the detection information acquired using the sensor 150. The management system 100 then communicates with a target vehicle in the parking lot 2 (for example, by wireless LAN) and performs AVP for the target vehicle. For example, consider a case where a user is attempting to enter the vehicle 1 using AVP. The user parks the vehicle 1 (target vehicle) to be entered at the vehicle stop position 2b in the entrance location 2a. After getting off the vehicle 1 (target vehicle) at the vehicle stop position 2b, the user makes a request for AVP. The management system 100 transmits command information such as a parking position and a driving route to the target vehicle based on management data related to the parking lot 2 and each vehicle 1. The AVP related to entry is performed by controlling the target vehicle according to the command information received by the control system of the target vehicle.

[0013] From the viewpoint of reliability and safety, AVP requires that the target vehicle be identified when entering the parking lot (identification can also be called authentication). This is to ensure that the vehicle 1 parked at the vehicle stop position 2b is the vehicle 1 (target vehicle) of the legitimate user. If vehicle identification is successful, the vehicle 1 parked at the vehicle stop position 2b is determined to be a legitimate target vehicle, and AVP is started. By identifying the target vehicle in this way, it is possible to allow only legitimate target vehicles to enter the parking lot. Conversely, if the target vehicle is not identified, there is a risk that unrelated vehicles that are not target vehicles will enter the parking lot. This is problematic from the viewpoint of reliability and safety. Similarly, in functions that provide control and services to other target vehicles (hereinafter referred to as "AVP and other functions"), it is usually required to identify the target vehicle.

[0014] The vehicle identification system according to this embodiment can be configured to identify a target vehicle in a function such as AVP. The vehicle identification system according to this embodiment employs a method of instructing the vehicle 1 to execute an action and identifying the target vehicle based on whether or not the action recognized by the sensor 150 matches an expected action (hereinafter referred to as an "action-based identification method").

[0015] Here, an action can be defined by a combination of a device (means) that executes the action and an operation pattern of the device. Devices that execute an action include lighting devices (headlights, brake lights, fog lights, etc.), interior lighting devices, horns, turn signals, windshield wipers, doors, door windows, engines, electric motors, etc. For example, one action is to flash headlights in a specified pattern. Another action is to sound a horn in a specified pattern. An action may include a request for repetition of an operation pattern.

[0016] The action-based identification technique requires the use of the sensor 150 to detect the action of the vehicle 1. For this reason, the sensor 150 is configured to be able to detect at least the action to be performed by the vehicle 1. For example, when a visually detectable visible action (such as an action related to the lighting of a lighting device) is to be performed, the sensor 150 is configured with a camera. Also, for example, when an audible action (such as an action related to the sounding of a horn) is to be performed, the sensor 150 is configured with a microphone.

[0017] The action-based identification method has the advantage of allowing freedom in designing the action used for vehicle identification and being highly flexible. On the other hand, the action-based identification method requires the vehicle 1 to execute an action having a certain duration, so vehicle identification takes a certain amount of time. For this reason, in the conventional technology (for example, the technology disclosed in Patent Document 1), it is assumed that the vehicle 1 continues to stop at the vehicle stop position 2b for a while from when vehicle identification starts to when it is completed. However, it is undesirable for the vehicle 1 to occupy a certain space for a long period of time, as this hinders efficient traffic. For example, in the AVP, this may hinder the efficient entry of the vehicle 1 and cause congestion of the vehicles 1 waiting to enter the parking lot.

[0018] In response to the above-mentioned problem, the vehicle identification system according to the present embodiment makes it possible to identify a target vehicle while suppressing the occurrence of situations that impede efficient traffic.

[0019] 2. Vehicle Identification System 2 is a conceptual diagram for explaining an overview of the vehicle identification system according to this embodiment. The vehicle identification system according to this embodiment can be configured with a management system 100 that executes processing, and a sensor 150 that detects an action executed by a vehicle 1.

[0020] In the vehicle identification system according to this embodiment, a vehicle identification zone 3 having a certain range is provided in front of a vehicle stop position 2b. The vehicle stop position 2b is a position where a target vehicle is requested to stop in functions such as AVP. It is assumed that a vehicle 1 from which a user wishes to receive control / services will pass through the vehicle identification zone 3, reach the vehicle stop position 2b, and stop there. The vehicle identification zone 3 may be given as a virtual range for information processing.

[0021] In the vehicle identification system according to this embodiment, vehicle identification begins when vehicle 1 (first vehicle) travels through vehicle identification zone 3. First, the management system 100 instructs vehicle 1 to start performing a predetermined action while vehicle 1 is traveling through vehicle identification zone 3. If vehicle 1 is a genuine target vehicle, vehicle 1 is expected to start performing the predetermined action in accordance with the instruction. Vehicle 1 may be configured to perform the predetermined action multiple cycles.

[0022] Next, the management system 100 starts a process (recognition process) of recognizing an action performed by the vehicle 1 (first vehicle) using the sensor 150 while the vehicle 1 (first vehicle) is traveling in the vehicle identification zone 3. Then, the management system 100 identifies the vehicle 1 performing the predetermined action as a target vehicle. For example, if the predetermined action is headlight flashing, the management system 100 recognizes the flashing pattern from the detection information of the sensor 150 in the recognition process. Then, when the recognized pattern matches the operation pattern of the predetermined action, the management system 100 determines that the predetermined action has been performed and identifies the vehicle 1 as a target vehicle. In particular, the vehicle identification system may be configured to complete the identification of the target vehicle before the vehicle 1 reaches the vehicle stop position 2b. For example, the management system 100 is configured to determine a predetermined action that indicates the duration of the action as an index according to the state (speed, position, etc.) of the vehicle 1.

[0023] After completing the identification of the target vehicle, the management system 100 starts providing control and services to the target vehicle stopped at the vehicle stop position 2b.

[0024] As described above, according to this embodiment, an instruction to start execution of a predetermined action is given and recognition processing is started before the vehicle 1 reaches the vehicle stop position 2b. This makes it possible to complete identification of the target vehicle or to recognize part of the action to be performed by the vehicle 1 when the vehicle 1 reaches the vehicle stop position 2b. Therefore, it is possible to shorten the time that the target vehicle stops at the vehicle stop position 2b in relation to vehicle identification. This in turn makes it possible to suppress the occurrence of a situation in which efficient traffic is impeded.

[0025] 3. Camera selection process In the following, in the vehicle identification system according to the present embodiment, a case will be considered in which the specified action to be instructed is a visible action, and the sensor 150 used in the recognition process is configured with a camera. In this case, the vehicle identification system needs to use a camera to recognize the visible action performed by the vehicle 1.

[0026] Incidentally, in functions such as AVP, it is assumed that the sensor 150 includes a plurality of cameras installed in a predetermined area (for example, parking lot 2). When applying the vehicle identification system according to this embodiment, a camera to be used for recognition processing is selected from the plurality of cameras. Then, while the vehicle 1 is traveling in the vehicle identification zone 3, the recognition processing using the selected camera is started.

[0027] Here, we consider how to select a camera to be used in the recognition process. A camera has a range in which it can accurately detect a visible action (hereinafter referred to as the "effective recognition range"), depending on its placement and specifications. The effective recognition range does not necessarily match the angle of view of the camera. The effective recognition range may also depend on the relative distance to the vehicle 1 to be recognized, the relationship between the orientation of the camera and the orientation of the vehicle 1 to be recognized, etc. In order to recognize the execution of a specified visible action by vehicle 1, it is necessary for vehicle 1 to be located within the effective recognition range of the camera to be used for the recognition process from the start to the end of the specified visible action.

[0028] From one viewpoint, since the vehicle identification zone 3 is a range that can be determined in advance, it is conceivable to select one camera to be used for the recognition process in advance. For example, it is conceivable to select a camera whose effective recognition range is a certain range from the starting position of the vehicle identification zone 3 as the camera to be used for the recognition process. However, the vehicle 1 moves a certain distance during the time from the start to the end of a predetermined visible action. For this reason, there is a possibility that the effective recognition range of a single camera cannot capture the entirety of the predetermined visible action from the start to the end. For example, this may occur when the vehicle 1 is traveling at a high speed or when the start of the predetermined visible action is delayed.

[0029] Fig. 3 is a conceptual diagram showing an example of the above problem. In Fig. 3, three cameras 151a, 151b, and 151c are shown as the sensor 150. Also shown are effective recognition ranges 152a, 152b, and 152c of the three cameras 151a, 151b, and 151c. The camera 151b is installed at a position to image the vehicle 1 after the camera 151a. Furthermore, the camera 151c is installed at a position to image the vehicle 1 after the camera 151b.

[0030] FIG. 3 shows a case where the camera 151a is a camera used for the recognition process. In FIG. 3, two patterns, pattern A and pattern B, are shown for the situation of the recognition process. Pattern A shows a situation that is normally expected. In pattern A, a section 4 from the start to the end of a predetermined visible action is included in the effective recognition range 152a of the camera 151a. Therefore, the execution of the predetermined visible action by the vehicle 1 can be normally recognized. On the other hand, pattern B shows a situation in which the start of the predetermined visible action is delayed. For example, a communication delay, a processing delay, or the like occurs. In pattern B, a part of section 4 is not included in the effective recognition range 152a. Therefore, the execution of the predetermined visible action by the vehicle 1 cannot be recognized, and vehicle identification fails.

[0031] In this way, when the execution of a predetermined visible action cannot be normally recognized by the default camera, it is possible to switch the camera used for the recognition process and resume the recognition process with the execution of the predetermined visible action in the next cycle. However, at this time, the recognition process performed by the default camera becomes a wasteful process. This is undesirable in terms of processing efficiency. Furthermore, if the camera is switched in response to the failure to recognize normally, there is a risk that the start of the recognition process using the switched camera will be delayed. As a result, there is a risk that the failure to identify the vehicle will be repeated. Note that it is also possible to secure a sufficient effective recognition range by selecting multiple cameras to be used in the recognition process in advance. However, at this time, there is a possibility that a camera will not be used in the recognition process as a result. This is also undesirable in terms of processing efficiency.

[0032] Therefore, the vehicle identification system of this embodiment can be configured to execute a process (camera selection process) to select a camera to be used in the recognition process prior to starting the recognition process, so as to be able to recognize the period from the start to the end of a specified visible action (hereinafter referred to as the "action section").

[0033] Fig. 4 is a conceptual diagram for explaining an overview of the camera selection process. In the camera selection process, a camera to be used in the recognition process is first provisionally selected. At the provisionally selected stage, the recognition process is not started. The camera provisionally selected first here may be a predetermined default camera. In Fig. 4, camera 151a is provisionally selected.

[0034] Next, in the camera selection process, a process (determination process) is performed to determine whether or not an action section can be recognized by a recognition process using a provisionally selected camera (provisional camera). The determination process can be performed by verifying whether or not a predicted section 4a of the action section is included in the effective recognition range of the provisional camera. The predicted section 4a can be calculated based on the position where a predetermined visible action is started (action start position), the duration of the predetermined visible action, and the speed of the vehicle 1. The action start position can be detected using the provisional camera. For example, the provisional camera is used to detect the rise (first shot) of a pattern of a predetermined visible action, and the position of the vehicle 1 at that time is obtained as the action start position. The duration of the predetermined visible action is information known in the management system 100. The speed of the vehicle 1 can be recognized using the provisional camera. Alternatively, it may be obtained by communication with the vehicle 1.

[0035] If the determination process determines that the action section can be recognized by the temporary camera, the temporary camera is selected as the camera to be used for the recognition process. On the other hand, if the action section cannot be recognized by the temporary camera, the temporary selection is redone and the temporary camera is changed. Then, the determination process is repeated again. In this way, a camera that has been confirmed in advance to be able to recognize the action section can be selected as the camera to be used for the recognition process.

[0036] Here, there are two examples of a method for changing the temporary camera. In the following examples, the temporary camera that was previously temporarily selected is called the "first camera." Also, one of the cameras installed in the vehicle identification zone 3 at a position to image the vehicle 1 after the first camera is called the "second camera." For example, in FIG. 4, when camera 151a is the first camera, the second camera is camera 151b or camera 151c.

[0037] In the first example (see FIG. 4), the second camera is provisionally selected as the provisional camera. In FIG. 4, camera 151a is the first camera, and camera 151b is provisionally selected as the provisional camera. The effective recognition range of the second camera is expected to include all action sections of the next and subsequent cycles. In other words, when the second camera is selected as the camera to be used in the recognition process, it is assumed that the target of the recognition process is the action section of the next and subsequent cycles. Therefore, in the re-determination process in the first example, it is verified whether or not the prediction section 4a of the next and subsequent cycles is included in the effective recognition range of the second camera. At this time, the action start position may be predicted from the action start position acquired previously, or may be detected using the second camera.

[0038] In a second example (see FIG. 4), a combination of the first camera and the second camera is provisionally selected as a provisional camera. In FIG. 4, camera 151a is assumed to be the first camera, and a combination of camera 151a and camera 151b is provisionally selected as a provisional camera. The effective recognition range of the combination of the first camera and the second camera is an extension of the effective recognition range of the first camera, and is therefore expected to include the prediction interval 4a. The re-determination process in the second example may be performed in relation to the previously calculated prediction interval 4a and the effective recognition range of the combination of the first camera and the second camera.

[0039] By employing the first or second example, the situation in which the determination process is excessively repeated is suppressed, so that the camera selection process can be performed effectively.

[0040] By executing the camera selection process as described above, the vehicle identification system can start the recognition process using a camera that has been confirmed in advance to be capable of recognizing action sections. This makes it possible to prevent failure of the vehicle identification process in the recognition process, and ultimately improves processing efficiency.

[0041] Fig. 5 is a flowchart showing an example of the process executed in the vehicle identification system with respect to the above-mentioned camera selection process. The process according to the flowchart shown in Fig. 5 is started after the vehicle 1 enters the vehicle identification zone 3 and communication between the management system 100 and the vehicle 1 is established, for example.

[0042] In step S100, the management system 100 instructs the vehicle 1 to start executing a predetermined visual action.

[0043] Next, in step S110, the management system 100 provisionally selects a provisional camera. The provisional selection in step S110 is the first provisional selection (initial selection) in the camera selection process. For example, the management system 100 provisionally selects a predetermined default camera as the provisional camera.

[0044] After step S110, the management system 100 performs a judgment process for the provisionally selected temporary camera. The management system 100 acquires an action start position using the temporary camera (step S120), and judges whether or not the temporary camera can recognize a predetermined visible action performed by the vehicle 1 from start to end (action section) (step S130).

[0045] When it is determined by the determination process that the action section cannot be recognized by the temporary camera (step S140; No), the management system 100 changes the temporary camera (step S150) and performs the determination process again. Here, when referring to the action start position acquired previously, the process related to step S120 can be skipped.

[0046] When it is determined by the determination process that the action section can be recognized by the temporary camera (step S140; Yes), the management system 100 selects the temporarily selected temporary camera as a camera to be used in the recognition process (step S160). Then, the management system 100 starts the recognition process using the selected camera (step S170).

[0047] 4. Hardware Configuration 6 is a diagram showing an example of a hardware configuration of a vehicle identification system 10 according to this embodiment. The vehicle identification system 10 is configured between a vehicle 1 and a management system 100 that communicates with the vehicle 1. The vehicle identification system 10 includes a sensor 150 that is communicatively connected to the management system 100.

[0048] The management system 100 includes a processor 110, a storage device 120, and a communication interface 130.

[0049] The processor 110 executes various processes. The processor 110 is configured with, for example, a central processing unit (CPU), a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), etc.

[0050] The storage device 120 is connected to the processor 110 and stores various information necessary for the processing executed by the processor 110. The storage device 120 is configured with a recording medium such as, for example, a random access memory (RAM), a read only memory (ROM), a solid state drive (SSD), or a hard disk drive (HDD).

[0051] The storage device 120 stores a computer program 121 and management data 122 .

[0052] The computer program 121 describes the processing to be executed by the processor 110. The processor 110, which executes the processing related to the computer program 121, and the storage device 120 cooperate with each other to realize information processing by the management system 100. In particular, information processing related to vehicle identification according to the present embodiment may be realized.

[0053] The management data 122 records various types of management information. For example, the management data 122 records management information related to a predetermined action (a device that executes the action, a motion pattern, etc.), management information related to the vehicle identification zone 3 (position, range, etc.), management information related to the sensor 150 (location, specifications, a default camera, etc.), etc.

[0054] The communication interface 130 is an interface for communicating with devices external to the management system 100 to transmit and receive information. For example, the communication interface 130 is composed of a device for connecting to surrounding devices via wireless LAN, a device for connecting to a mobile communication network, a device for connecting to the Internet, etc. The management system 100 transmits and receives information to and from the vehicle 1 via the communication interface 130.

[0055] The vehicle 1 includes a control processing device 210, an on-vehicle sensor 220, a communication interface 230, and an on-vehicle device 240.

[0056] The control processing device 210 is communicatively connected to the on-board sensor 220, the communication interface 230, and the on-board device 240. The control processing device 210 is a computer that performs information processing related to the control of the vehicle 1 based on various information. The control processing device 210 may include a processor and a storage device. The information processing by the control processing device 210 may be realized by cooperation between the processor and the storage device.

[0057] For example, the control processing device 210 is configured with one or more ECUs (Electronic Control Units). Also, for example, the control processing device 210 is configured with a kit (for example, an AVP kit) for functions provided by the management system 100. The control processing device 210 generates and outputs a control signal for controlling the vehicle 1 by information processing. In particular, when the vehicle 1 receives an instruction from the management system 100 to start executing a predetermined action, the control processing device 210 generates and outputs a control signal for executing the predetermined action. The control signal is transmitted to the in-vehicle device 240.

[0058] The on-board sensor 220 detects information on the surrounding environment and driving conditions of the vehicle 1. Examples of the on-board sensor 220 include a camera, a radar, a LiDAR, a wheel speed sensor, an IMU (Inertial Measurement Unit), a GNSS (Global Navigation Satellite System) sensor, and the like.

[0059] The communication interface 230 is an interface for communicating with devices external to the vehicle 1 to transmit and receive information. The vehicle 1 transmits and receives information to and from the management system 100 via the communication interface 230.

[0060] The in-vehicle device 240 includes lighting devices, interior lighting devices, a horn, direction indicators, wipers, doors, door windows, driving devices, braking devices, steering devices, etc. Each device of the in-vehicle device 240 includes an actuator 241 that can be controlled by the control processing device 210. The in-vehicle device 240 acquires a control signal from the control processing device 210. The actuator 241 operates according to the acquired control signal, whereby the in-vehicle device 240 is controlled by the control processing device 210. Also, the control of the vehicle 1 is realized by the control of the in-vehicle device 240. [Explanation of symbols]

[0061] 1 vehicle, 3 vehicle identification zones, 10 vehicle identification systems, 100 management systems, 110 processor, 120 storage device, 121 computer program, 122 management data, 150 sensors

Claims

1. A vehicle identification system for identifying a target vehicle, comprising: one or more processors; The one or more processors: instructing the first vehicle to initiate execution of a predetermined action to identify the target vehicle while the first vehicle is traveling through the vehicle identification zone; while the first vehicle is traveling in the vehicle identification zone, starting a recognition process for recognizing an action performed by the first vehicle using a sensor; Identifying the first vehicle that performs the predetermined action as the target vehicle. It is configured as follows: Vehicle Identification System.

2. 2. The vehicle identification system of claim 1, the predetermined action is a predetermined visible action; The recognition process includes recognizing a visible action performed by the first vehicle using at least one camera; The one or more processors further include: A camera selection process is performed to select the at least one camera so that the predetermined visible action performed by the first vehicle can be recognized from start to end. It is configured as follows: Vehicle Identification System.

3. 3. The vehicle identification system according to claim 2, The camera selection process includes: Temporarily selecting a temporary camera; a determination process for determining whether the predetermined visible action performed by the first vehicle can be recognized by the virtual camera from start to finish; selecting the virtual camera as the at least one camera if the predetermined visible action can be recognized by the virtual camera from start to finish; Includes Vehicle Identification System.

4. 4. The vehicle identification system according to claim 3, The second camera is installed at a position behind the first camera so as to capture an image of the first vehicle, The camera selection process includes: tentatively selecting the first camera as the tentative camera and performing the determination process; when the first camera cannot recognize the predetermined visible action from the start to the end, the second camera is provisionally selected as the provisional camera instead of the first camera, and the determination process is performed; Includes Vehicle Identification System.

5. 4. The vehicle identification system according to claim 3, The second camera is installed at a position behind the first camera so as to capture an image of the first vehicle, The camera selection process includes: tentatively selecting the first camera as the tentative camera and performing the determination process; when the first camera cannot recognize the predetermined visible action from the start to the end, a combination of the first camera and the second camera is provisionally selected as the provisional camera, and the determination process is performed; Includes Vehicle Identification System.

Images