JP6885355B2 - Identification device, identification method, and identification program - Google Patents

Description

The present invention relates to a technique for associating a vehicle captured with images captured at two different points.

Conventionally, there has been a device for acquiring the travel time of a vehicle between two points on a vehicle's travel path (general road, highway, etc.) as traffic information (see, for example, Patent Document 1). Patent Document 1 has a configuration in which the travel time at the A point and the B point is acquired as follows.

The characteristics of each vehicle (vehicle size, vehicle color) are extracted for the vehicle group that has passed point A, and the characteristics of each vehicle (vehicle size, vehicle color) are also extracted for the vehicle group that has passed point B. ) Is extracted. When a vehicle group having the same or similar characteristics as the characteristics of each vehicle extracted for the vehicle group that has passed the A point is detected at the B point, this vehicle group is determined to be the vehicle group that has passed the A point. Then, the elapsed time from the time when this vehicle group passed the point A (the time detected at the point A) to the time when the vehicle group reached the point B (the time detected at the point B) is defined as the point A. Obtained as travel time with point B.

Further, Patent Document 1 also describes a configuration in which the license plate of a vehicle is read and the travel time between the two points is acquired by associating the vehicles at the two points with the license plate.

JP-A-2009-129278

However, in the system disclosed in Patent Document 1, it is necessary to attach an imaging device (camera) at a high place in order to image a vehicle traveling on a traveling road from above. For this reason, it was necessary to erect a pole or the like for attaching the imaging device. In addition, the work related to the installation and maintenance of the image pickup device becomes work at a high place. Therefore, there is a problem that the cost for constructing and operating the system increases.

An object of the present invention is to provide a technique capable of associating vehicles between two points while suppressing the cost of constructing and operating the system.

The identification device of the present invention is configured as follows in order to achieve the above object.

The image input unit includes a first point image in which the wheels of a vehicle traveling on the road at the first point are imaged, and a second image of the wheels of a vehicle traveling on the road at the second point. The point image is input. The first point and the second point are different points. For example, the first point image and the second point image are images obtained by capturing the wheels of the vehicle in the width direction of the traveling path.

The feature amount extraction unit processes the first point image and the second point image to extract the feature amount of the wheel. Vehicle wheels include wheels and tires. The wheels are attached to the axles of the vehicle with bolts and nuts. The tires are attached to the wheels. Wheels vary in their outer diameter, the number of bolts used to attach them to the axle of the vehicle, the shape of the vents, the pattern drawn on the center cap, and so on. That is, there are many types of wheels attached to the vehicle. Therefore, the wheel feature amount (mainly the wheel feature amount) can be used as the feature amount for identifying the vehicle. Then, the corresponding portion includes a vehicle in which the wheel is imaged in the first point image and a vehicle in which the wheel is imaged in the second point image by collating the feature amount of the wheel extracted by the feature amount extraction unit. Correspond.

Therefore, the image pickup device such as a camera may be installed in a place where the wheels of the vehicle traveling on the traveling road can be imaged. In other words, the image pickup device such as a camera can be installed at a height of about several tens of centimeters to 1 m with respect to the road surface of the traveling path on which the vehicle travels. In this way, since the image pickup device such as a camera can be installed in a relatively low place, it is not necessary to erect a pole or the like for attaching the image pickup device such as a camera, and the image pickup device such as a camera is attached and maintained. The work required for work does not become work at height. Therefore, the cost for constructing and operating the travel time measurement system can be sufficiently suppressed.

In addition, the feature amount extraction unit performs resizing correction for converting the wheels captured in the first point image and the second point image to the reference size, and the feature amount of the wheels is obtained with respect to the image after the resizing correction. It is preferable to have a configuration for extracting. By performing this resizing correction, in extracting the outer diameter of the wheel (or the ratio with the outer diameter of the wheel), which is a feature amount, the influence of the difference in the distance between the image pickup device such as a camera and the wheel at the time of imaging can be suppressed. ..

In addition, the feature amount extraction unit performs tilt correction for the first point image and the second point image so that the imaging direction is aligned with the rotation axis of the wheel, and the tilt-corrected image of the wheel. It is preferable to configure the structure to extract the feature amount. By performing this tilt correction, the number of bolts, the shape of the vents, the pattern drawn on the center cap, etc., which are the feature quantities, are extracted due to the deviation between the image pickup axis of the image pickup device such as a camera and the axle. , The influence of distortion of the wheel image is suppressed.

In addition, the date and time of imaging are associated with the first point image and the second point image, and the correspondence between the vehicle in which the wheels are imaged in the first point image and the vehicle in which the wheels are imaged in the second point image in the identification unit. The travel time between the first point and the second point may be measured based on the attached result and the imaging date and time associated with the first point image and the second point image.

According to the present invention, the cost for constructing and operating the system can be suppressed.

It is a figure which shows the structure of the travel time measurement system. It is a schematic view which shows the travel path of the vehicle which the travel time is measured by the travel time measurement system, FIG. 2 (A) is a plan view seen from above of the travel path, and FIG. 2 (B) is a travel path. It is a plan view seen in the width direction of. It is a block diagram which shows the structure of the main part of a point device. It is a block diagram which shows the structure of the main part of an identification apparatus. It is a figure which shows the collation data. 6 (A) and 6 (B) are views showing the wheels of a vehicle. It is a flowchart which shows the operation of a point device. It is a flowchart which shows the operation of the identification apparatus. It is the schematic which shows the travel path of the vehicle which the travel time is measured by the travel time measurement system which concerns on another example.

Hereinafter, embodiments of the present invention will be described.

<1. Application example>
FIG. 1 is a diagram showing a configuration of a travel time measurement system to which the identification device according to the embodiment of the present invention is applied. FIG. 2 is a schematic view showing a travel path of a vehicle for which travel time is measured by this travel time measurement system. FIG. 2A is a plan view seen from above the running path, and FIG. 2B is a plan view seen in the width direction of the running path. In this travel time measurement system, a point device 5, a camera 6, and a vehicle detector 7 are installed at each point that becomes a starting point or an ending point for measuring the travel time. FIG. 2 illustrates a first point and a second point. The camera 6 and the vehicle detector 7 are connected to the point device 5. Further, the point device 5 at each point is connected to the identification device 1 installed in the center or the like via the network 2 so as to be capable of data communication.

The camera 6, which is an imaging device, images the wheels of a vehicle 100 traveling on a traveling road (FIG. 2, shows three vehicles 100a, 100b, and 100c). In other words, the imaging region of the camera 6 is set to a height at which the wheels of the vehicle 100 traveling on the traveling road pass. The camera 6 is installed at a height of about several tens of centimeters to 1 m with respect to the road surface of the traveling path. Further, the camera 6 is installed so that the image pickup axis is substantially parallel to the width direction of the traveling path. The camera 6 is installed in the roadside zone of the traveling road and does not become an obstacle to the vehicle 100 traveling on the traveling road.

The vehicle detector 7 detects the presence or absence of the vehicle 100 at the vehicle detection position. The vehicle detection position is set on the upstream side of several tens of centimeters of the imaging area of the camera 6. The vehicle detector 7 may be configured to detect a vehicle by using a known photoelectric sensor, ultrasonic sensor, loop coil type vehicle detection sensor, or the like. The vehicle detector 7 outputs a vehicle detection signal (a signal indicating the presence or absence of a vehicle) to the point device 5.

Further, the point device 5 outputs a release signal to the camera 6 when the vehicle detection signal input from the vehicle detector 7 changes from no vehicle to with vehicle. The camera 6 captures a plurality of frames (about 2 to 5 frames) of the imaging area in response to the input of the release signal from the point device 5. The camera 6 outputs the frame image captured this time to the point device 5. The point device 5 extracts a frame image in which the wheels of the vehicle 100 are captured from a plurality of frame images input from the camera 6, and transmits the extracted frame image to the identification device 1. The point device 5 associates the frame image with the imaging date and time and transmits the frame image to the identification device 1.

The identification device 1 stores the feature amount of the wheel extracted by processing the frame image transmitted from the point device 5 at that point in association with the imaging date and time of the frame image for each point. The identification device 1 associates the vehicle 100 with two points for measuring the travel time and measures the travel time. The association of the vehicle 100 is performed by collating the feature amounts of the wheels extracted from the captured image.

As described above, in the travel time measurement system according to this example, the association of the vehicle 100 at two different points is performed by the feature amount of the wheel extracted by processing the frame image of the wheel of the vehicle 100. Therefore, since the camera 6 only needs to be able to image the wheels of the vehicle 100 traveling on the traveling road, the camera 6 can be installed at a height of about several tens of centimeters to 1 m with respect to the road surface of the traveling road on which the vehicle 100 travels. That is, the camera 6 can be installed in a relatively low place. Therefore, it is not necessary to erect a pole or the like for attaching the camera 6, and the work related to the attachment and maintenance of the camera 6 is not a high-place work. Therefore, the cost for constructing and operating the travel time measurement system can be sufficiently suppressed.

The processing related to the image in each device is the processing for the image data.

<2. Configuration example>
FIG. 3 is a block diagram showing a configuration of a main part of a point device in the travel time measurement system according to this example. The point device 5 includes a control unit 21, a detection signal input unit 22, a camera connection unit 23, an image storage unit 24, and a communication unit 25. The point device 5 is installed at a point that becomes a starting point or an ending point for measuring travel time.

The control unit 21 controls the operation of each unit of the point device 5 main body. The control unit 21 is composed of a hardware CPU, a memory, and other electronic circuits. The hardware CPU executes a program for controlling the operation of each part of the point device 5 main body. In addition, the memory has an area for expanding the program and an area for temporarily storing data and the like generated when the program is executed. The control unit 21 may be an LSI in which a hardware CPU, a memory, and the like are integrated.

A vehicle detector 7 is connected to the detection signal input unit 22. A vehicle detection signal that detects the presence or absence of the vehicle 100 at the vehicle detection position is input to the detection signal input unit 22 from the vehicle detector 7.

The camera 6 is connected to the camera connection unit 23. The camera 6 captures a plurality of frame images when a release signal is input. That is, the camera 6 may have a continuous shooting function for continuously shooting and capturing still images, and may not have a function for capturing moving images. The camera connection unit 23 outputs a release signal to the camera 6. Further, the camera 6 outputs a plurality of frame images captured in response to the input release signal to the camera connection unit 23.

The point device 5 extracts one frame image from the plurality of frame images input to the camera connection unit 23. Specifically, the point device 5 extracts a frame image in which the entire wheel of the vehicle 100 is captured from a plurality of frame images input to the camera connection unit 23. Further, in the point device 5, if there are a plurality of frame images in which the entire wheel of the vehicle 100 is imaged, the imaged position of the wheel on the frame image is the frame closest to the center of the traveling direction of the vehicle 100 on the frame image. Extract the image (frame image with the least wheel strain). Further, the point device 5 associates the frame image extracted here with the imaging date and time of the frame image. The imaging date and time may be, for example, the time when the release signal is output to the camera 6, the time when the frame image is input from the camera 6, or the time when the detection signal input unit 22 is input. It may be the time when the vehicle detection signal changes from no vehicle to with vehicle. Further, if the camera 6 outputs the frame image in association with the imaging date and time, the imaging date and time associated with the frame image input from the camera 6 may be used as it is.

The image storage unit 24 stores the frame image extracted from the plurality of frame images input to the camera connection unit 23 in association with the imaging date and time of the frame image.

The communication unit 25 performs data communication with the identification device 1 via the network 2. The point device 5 transmits the frame image of the wheel of the vehicle 100 to the identification device 1 in association with the imaged date and time.

FIG. 4 is a block diagram showing the configuration of the main part of the identification device. The identification device 1 includes a control unit 11, a communication unit 12, a feature amount storage database 13 (feature amount storage DB 13), and an output unit 14.

The control unit 11 controls the operation of each part of the identification device 1 main body. Further, the control unit 11 has a feature amount extraction unit 11a, a corresponding attachment unit 11b, and a travel time measurement unit 11c. Details of the feature amount extraction unit 11a, the corresponding attachment unit 11b, and the travel time measurement unit 11c will be described later.

The communication unit 12 performs data communication with the point device 5 installed at each point via the network 2. As described above, the identification device 1 receives the frame image of the wheels of the vehicle 100 transmitted from the point device 5 in the communication unit 12. The frame image received by the communication unit 12 is associated with the imaging date and time. The communication unit 12 corresponds to the image input unit referred to in the present invention.

The feature amount storage DB 13 stores the collation data of the vehicle 100 for each point. FIG. 5 shows collation data for the first point. This collation data is registered by associating the imaging date and time, the image data, and the feature amount for each frame image in which the wheels of the vehicle 100 are imaged, which is transmitted from the point device 5 at that point. The details of the feature amount will be described later.

The output unit 14 outputs the measured travel time between the two points to a higher-level device such as a control center.

Next, the feature amount extraction unit 11a, the corresponding attachment unit 11b, and the travel time measurement unit 11c included in the control unit 11 will be described.

The feature amount extraction unit 11a processes the frame image in which the wheel of the vehicle 100 is imaged transmitted from the point device 5, and extracts the feature amount of the imaged wheel. 6 (A) and 6 (B) are views showing the wheels of a vehicle. The wheels of the vehicle 100 have wheels 50 and tires 60. The wheel 50 is attached to the axle of the vehicle 100 with bolts 51 and nuts (not shown). The tire 60 is attached to the wheel 50. The outer diameter of the wheel 50, the number of bolts 51 used to attach it to the axle of the vehicle 100, the shape of the ventilation holes 52, the pattern drawn on the center cap 53, and the like are various. That is, there are many types of wheels 50 attached to the vehicle 100. For example, the wheel 50 shown in FIG. 6A has six bolts 51 used to be attached to the axle of the vehicle 100, and the shape of the ventilation holes 52 is a triangle with rounded corners. The pattern drawn on the center cap 53 is a star. Further, the wheel 50 shown in FIG. 6B has five bolts 51 used for attaching to the axle of the vehicle 100, the shape of the ventilation holes 52 is circular, and is drawn on the center cap 53. The pattern is an airplane.

The feature amount extraction unit 11a has a frame image in which the wheels of the vehicle 100 are imaged, which is transmitted from the point device 5, and the outer diameter of the wheels being imaged is set to a predetermined size (the present invention). Performs resizing correction processing to correct to (reference size referred to in). The outer diameter of the wheel referred to here is the outer diameter of the tire 60 attached to the wheel 50. Further, the feature amount extraction unit 11a corrects the inclination of the wheel of the vehicle 100 so that the outer diameter of the wheel of the vehicle 100 becomes substantially a perfect circle with respect to the frame image in which the wheel of the vehicle 100 is captured, which is transmitted from the point device 5. Perform processing.

The resizing correction process is a process for acquiring the outer diameter of the wheel 50 as a ratio to the outer diameter of the wheel. By performing this resizing correction process, the outer diameter of the wheel 50 (or the ratio to the outer diameter of the wheel) can be obtained by suppressing the influence of the difference in the distance between the camera 6 and the wheel at the time of imaging. The tilt correction process is a process for aligning the imaging direction of the camera 6 with the axle of the vehicle 100. That is, the tilt correction process is a process for obtaining an image of the wheels of the vehicle 100 captured from a direction parallel to the axle. By performing this tilt correction processing, it is possible to suppress the distortion of the wheel in the frame image, and it is possible to accurately detect the number of bolts 51, the shape of the ventilation holes 52, the features such as the pattern drawn on the center cap 53, and the like. ..

In this example, the feature amount extraction unit 11a processes the frame image in which the wheels of the vehicle 100 are captured, which are transmitted from the point device 5, the outer diameter of the wheels 50, the number of bolts 51, and the ventilation holes 52. The shape, the pattern drawn on the center cap 53, and the like are extracted as feature quantities. The feature amount extracted by the feature amount extraction unit 11a is registered as the feature amount of the collation data shown in FIG.

The feature amount extracted by the feature amount extraction unit 11a may be a part of the above-mentioned ones, or the feature amount may be extracted for other items as well.

For the two selected points, the corresponding unit 11b collates the feature amount of the matching data stored for one point with the feature amount of the matching data stored for the other point, and collates the same vehicle. Correspond to the frame image in which 100 wheels are imaged. That is, the corresponding portion 11b associates the vehicle 100 between two points by using the feature amount of the wheel 50.

The travel time measuring unit 11c uses the time difference between the imaging date and time of the wheels at one point and the imaging date and time of the wheels at the other point for the vehicle 100 associated between the two points by the corresponding attachment unit 11b. Measure the travel time between two points.

The control unit 11 of the identification device 1 is composed of a hardware CPU, a memory, and other electronic circuits. When the hardware CPU executes the identification program according to the present invention, it operates as the feature amount extraction unit 11a, the corresponding attachment unit 11b, and the travel time measurement unit 11c. Further, the memory has an area for developing the identification program according to the present invention and an area for temporarily storing data and the like generated when the identification program is executed. The control unit 11 may be an LSI in which a hardware CPU, a memory, and the like are integrated. Further, the hardware CPU is a computer that executes the identification method according to the present invention.

<3. Operation example>
Next, the operations of the point device 5 and the identification device 1 of this travel time measurement system will be described. FIG. 7 is a flowchart showing the operation of the point device. The point device 5 determines whether or not the vehicle detection signal input from the vehicle detector 7 to the detection signal input unit 22 has changed from the absence of the vehicle 100 to the presence of the vehicle 100 (s1), and a frame to the identification device 1. Whether or not the image transmission timing has come (s2) is repeatedly determined.

When the point device 5 determines that the vehicle detection signal input to the detection signal input unit 22 has changed to have a vehicle, it outputs a release signal to the camera 6 connected to the camera connection unit 23 (s3). .. When the release signal is input, the camera 6 captures a plurality of frame images and outputs the captured frame images to the point device 5.

As described above, the vehicle detection position where the vehicle detector 7 detects the presence or absence of the vehicle 100 is several tens of centimeters upstream of the image pickup area of the camera 6. Therefore, the point device 5 outputs a release signal to the camera 6 connected to the camera connection unit 23 at the timing when the vehicle detection signal input to the detection signal input unit 22 changes to have a vehicle. , The front wheel of the vehicle 100 can be imaged by the camera 6.

When a plurality of frame images are input from the camera 6 connected to the camera connecting portion 23, the point device 5 captures the entire wheel of the vehicle 100 from the plurality of frame images input this time1. One frame image is extracted (s4, s5). In the point device 5, if there are a plurality of frame images in which the entire wheel of the vehicle 100 is imaged in s5, the imaged position of the wheel on the frame image is closest to the center of the traveling direction of the vehicle 100 on the frame image. Extract the frame image.

The point device 5 associates the frame image extracted in s5 with the imaging date and time of this frame image and stores it in the image storage unit 24 (s6), and returns to s1. The imaging date and time associated with the frame image may be, for example, the time when the release signal is output to the camera 6, the time when the frame image is input from the camera 6, or the detection signal input unit 22 in s1. It may be the time when it is determined that the vehicle detection signal input to is changed to have a vehicle. Further, if the camera 6 outputs the frame image in association with the imaging date and time, the imaging date and time associated with the frame image input from the camera 6 may be used as it is.

Each time the processing of s3 to s6 described above is performed, the point device 5 stores the frame image and the imaging date and time of the frame image in the image storage unit 24 in association with each other.

Further, when the point device 5 determines in s2 that the transmission timing is reached, it is a frame image stored in the image storage unit 24 and has not been transmitted to the identification device 1 (a frame image that has not been transmitted). ) Is transmitted to the identification device 1 (s7), and the process returns to s1. In s7, the imaging date and time of the frame image is associated with each frame image transmitted to the identification device 1.

When the point device 5 transmits the frame image to the identification device 1 in s7, the frame image transmitted this time may be deleted from the image storage unit 24. With this configuration, the storage capacity of the image storage unit 24 can be suppressed. Further, it is preferable that the point devices 5 installed at each point are set so that the transmission timings related to s2 are dispersed.

Further, the transmission timing of the frame image in the point device 5 is preferably set to an interval of about several minutes to several minutes.

Next, the operation of the identification device 1 will be described. FIG. 8 is a flowchart showing the operation of the identification device. The identification device 1 repeatedly determines whether or not the frame image is received from the point device 5 (s11) and whether or not the travel time between the two points is measured (s12). As described above, the point device 5 transmits the frame image stored in the image storage unit 24 to the identification device 1 at the set transmission timing.

When the identification device 1 receives the frame image transmitted from the point device 5, the identification device 1 executes the processes related to s13 to s18 shown below for each received frame image. The feature amount extraction unit 11a of the identification device 1 selects a frame image to be processed from the received frame images (s13). The feature amount extraction unit 11a performs a cutting process for cutting out the wheels of the vehicle 100 imaged in the frame image selected in s13 (s14). The process applied to s14 suppresses the load of the process applied after s15. Specifically, the entire frame image transmitted from the point device 5 is not subjected to the processing after s15, but the image of the wheels of the vehicle 100 cut out in s14 is subjected to the processing after s15. By doing so, the processing load is suppressed. Therefore, the identification device 1 may be configured not to execute the process related to s14.

The feature amount extraction unit 11a enlarges or reduces the image of the vehicle 100 cut out in s14, and determines the outer diameter of the wheel (in this example, the outer diameter of the tire 60 attached to the wheel 50) in advance. Resize correction processing to match the reference size is performed (s15). In addition, the feature amount extraction unit 11a performs tilt correction processing for aligning the imaging direction with the axle of the vehicle 100 with respect to the image of the vehicle 100 cut out in s14 (s16). In the tilt correction process for s16, the tilt may be corrected so that the outer diameter of the tire 60 attached to the wheel 50 becomes a perfect circle, but the tire 60 of the traveling vehicle 100 is flat. Therefore, it is preferable to correct the inclination so that the outer diameter of the wheel 50 becomes a perfect circle.

The processing related to s15 and s16 is not limited to the above order, and may be performed in the reverse order.

The feature amount extraction unit 11a performs a feature amount extraction process for extracting the feature amount of the wheel 50 from the image of the wheel that has undergone the above-mentioned resizing correction process and the tilt correction process (s17). In this example, in the feature amount extraction process applied to s17, the outer diameter of the wheel 50, the number of bolts 51, the shape of the ventilation holes 52, and the pattern drawn on the center cap 53 are extracted as feature amounts. However, the outer diameter of the wheel 50 may be the ratio of the outer diameter of the wheel 50 to the outer diameter of the tire 60, or may be the length of the outer diameter of the wheel 50 on the image corrected for resizing.

The feature amount extraction unit 11a generates collation data for the frame image to be processed this time, and registers the generated collation data in the feature amount storage DB 13 (s18). As shown in FIG. 5, the collation data is associated with the imaging date and time, the image data, and the feature amount. The image data may be configured not to be stored. The feature amount extraction unit 11a determines whether or not the frame image received this time includes a frame image (unprocessed frame image) for which the above-mentioned processes related to s13 to s18 have not been executed (s19), and unprocessed. If there is a frame image of, the process returns to s13. Further, when the feature amount extraction unit 11a determines in s19 that there is no unprocessed frame image, the feature amount extraction unit 11a returns to s11.

Further, when the identification device 1 determines that the travel time is measured in s12, the identification device 1 determines between two points for measuring the travel time (s20). One point between the two points for measuring the travel time and the other point are the points where the above-mentioned camera 6 is installed. In other words, one point between the two points for measuring the travel time and the other point are points where the collation data is stored in the feature amount storage DB 13.

The corresponding unit 11b reads out the collation data of one point between the two points for measuring the travel time and the collation data of the other point between the two points for measuring the travel time from the feature amount storage DB 13, and the data thereof. An identification process for associating the vehicle 100 between two points using the feature amount is performed (s21). The identification process related to s21 is a known DP (Dynamic Programming) using not only the similarity of the feature amount of the wheel 50 registered in the feature amount storage DB 13 but also the passing order of the vehicle 100 at both points. The vehicle 100 is associated with each other by matching. When measuring the travel time between two points in substantially real time, the associating process for s25 is a point on the downstream side between the two points for measuring the travel time during the latest predetermined time (for example, 10 minutes). Each vehicle 100 that has passed through the above may be targeted and associated with the vehicle 100 that has passed the upstream side.

The travel time measuring unit 11c calculates the required time from passing the upstream side point to reaching the downstream side point for each vehicle 100 associated with the identification process related to s21, and sets the required time for each vehicle 100. Based on this, the travel time between two points is measured (s22). In s22, the average value of the time required from passing the upstream point calculated for each vehicle 100 to reaching the downstream point may be measured as the travel time between these two points, or the vehicle 100 may be measured. The median time required to reach the downstream point after passing the upstream point calculated for each may be measured as the travel time between these two points.

The identification device 1 determines whether to measure the travel time between the other two points (s23), and returns to s20 when it determines that the travel time is measured between the other two points. Further, the identification device 1 returns to s11 when it determines that there is no other location between the two points for measuring the travel time.

As described above, the travel time measurement system according to this example associates the vehicle 100 at two different points by using the feature amount of the wheel 50 of the vehicle 100. Therefore, the camera 6 that images the vehicle 100 installed at each point only needs to be able to image the wheel 50 of the vehicle 100, and is installed at a height of about several tens of cm to 1 m with respect to the road surface of the traveling path on which the vehicle 100 travels. it can. That is, the camera 6 can be installed in a relatively low place. Therefore, it is not necessary to erect a pole or the like for attaching the camera 6, and the work related to the attachment and maintenance of the camera 6 is not a high-place work. Therefore, the cost for constructing and operating the travel time measurement system can be sufficiently suppressed. In addition, the frequency with which the camera 6 can be installed and the camera 6 can be maintained without traffic regulation such as lane regulation increases. Therefore, it is possible to suppress the frequency of traffic congestion due to traffic regulations such as lane regulations.

Further, since the wheel 50 has various outer diameters, the number of bolts 51 used for attaching to the axle of the vehicle 100, the shape of the ventilation holes 52, the pattern drawn on the center cap 53, and the like, there are two points. Since the vehicles 100 can be associated with each other with high accuracy, the measurement accuracy of the travel time is not lowered.

<4. Modification example>
In the above example, the travel path of the vehicle 100 has been described with one lane, but as shown in FIG. 9, when the travel path of the vehicle 100 has two lanes, the point device 5, the camera 6, and the vehicle are located at each point. Two sets of detectors 7 may be installed. The two sets of point devices 5, the camera 6, and the vehicle detector 7 may be installed at positions substantially opposite to each other across the travel path. In this case, the vehicle detector 7 targets the vehicle 100 traveling in the lane near the installation position of the connected camera 6.

A plurality of sets of cameras 6 and vehicle detectors 7 may be connected to one point device 5.

Further, in the above example, the identification device 1 is provided with the feature amount extraction unit 11a, but the feature amount extraction unit 11a may be provided in the point device 5. In this case, the point device 5 may extract the feature amount of the wheel 50 and transmit the extracted feature amount to the identification device 1. With this configuration, the point device 5 does not need to transmit the frame image in communication with the identification device 1, so that the communication traffic can be significantly reduced.

Further, the camera 6 may be a video camera that captures a moving image. In this case, the camera 6 inputs the frame image related to the captured moving image to the point device 5. The point device 5 may be configured to extract a frame image in which the wheels of the vehicle 100 are captured by a known image processing technique such as pattern matching, and perform the processing related to s6 described above. With this configuration, the vehicle detector 7 can be eliminated.

The present invention is not limited to the above-described embodiment as it is, and at the implementation stage, the components can be modified and embodied within a range that does not deviate from the gist thereof. In addition, various inventions can be formed by an appropriate combination of the plurality of components disclosed in the above-described embodiment. For example, some components may be removed from all the components shown in the embodiments. In addition, components from different embodiments may be combined as appropriate.

Further, the correspondence between the configuration according to the present invention and the configuration according to the above-described embodiment can be described as the following appendix.
<Additional notes>
A first point image in which the wheels of the vehicle (100) are imaged at the first point and a second point image in which the wheels of the vehicle (100) are imaged at a second point different from the first point are input. Image input unit (12) and
The feature amount extraction unit (11a) that processes the first point image and the second point image to extract the feature amount of the wheel, and the feature amount extraction unit (11a).
By collating the feature amounts of the wheels extracted by the feature amount extraction unit (11a), the vehicle (100) in which the wheels are imaged in the first point image and the vehicle in which the wheels are imaged in the second point image. An identification device (1) including a corresponding attachment portion (11b) for associating with (100).

1 ... Identification device 2 ... Network 5 ... Point device 6 ... Camera 7 ... Vehicle detector 11 ... Control unit 11a ... Feature amount extraction unit 11b ... Corresponding part 11c ... Travel time measurement unit 12 ... Communication unit 13 ... Feature amount storage database (Feature quantity storage DB)
14 ... Output unit 21 ... Control unit 22 ... Detection signal input unit 23 ... Camera connection unit 24 ... Image storage unit 25 ... Communication unit 50 ... Wheel 51 ... Bolt 52 ... Vent hole 53 ... Center cap 60 ... Tire 100 (100a, 100b) , 100c) ... Vehicle

Claims (6)

An image of the first point, which is an image of the wheels of a vehicle traveling on the driving road at the first point, and an image of the wheels of a vehicle traveling on the traveling road at a second point different from the first point are imaged. The image input section where the second point image is input and
A feature amount extraction unit that processes the first point image and the second point image to extract the feature amount of the wheel, and the feature amount extraction unit.
By collating the feature amount of the wheel extracted by the feature amount extraction unit, the corresponding part that associates the vehicle in which the wheel is imaged in the first point image with the vehicle in which the wheel is imaged in the second point image. And with
The feature amount extraction unit performs resizing correction for converting the wheel imaged in the first point image and the second point image into a reference size, and the size-corrected image is obtained from the wheel. An identification device that extracts the feature amount of the wheel as the feature amount. The identification device according to claim 1, wherein the first point image and the second point image are images obtained by capturing the wheels of a vehicle in the width direction of a travel path. The feature amount extraction unit performs tilt correction for the first point image and the second point image so that the imaging direction is aligned with the axle of the wheel, and the tilt-corrected image is for the wheel. The identification device according to any one of claims 1 or 2 , which extracts the feature amount of the above. An imaging date and time is associated with the first point image and the second point image.
Correlation result, and the first point image and the second point image of the wheel on the vehicle and the second point image wheels on the first point image is imaged in association unit is captured vehicle based on the shooting date and time associated with the, with a travel time measuring unit for measuring the travel time between the second point and the first point, according to any of claims 1 to 3 Identification device. An image of the first point, which is an image of the wheels of a vehicle traveling on the running path at the first point, input to the image input unit, and running on the running path at a second point different from the first point. A feature amount extraction step that extracts the feature amount of the wheel by processing the second point image that images the wheel of the vehicle
By collating the feature amount of the wheel extracted in the feature amount extraction step, the corresponding step of associating the vehicle in which the wheel is imaged in the first point image with the vehicle in which the wheel is imaged in the second point image. And the computer runs ,
In the feature amount extraction step, resizing correction is performed to convert the wheels imaged in the first point image and the second point image into a reference size, and the resizing-corrected image of the wheels is subjected to resizing correction. An identification method , which is a step of extracting the feature amount of the wheel as the feature amount. An image of the first point, which is an image of the wheels of a vehicle traveling on the running path at the first point, input to the image input unit, and running on the running path at a second point different from the first point. A feature amount extraction step that extracts the feature amount of the wheel by processing the second point image that images the wheel of the vehicle
By collating the feature amount of the wheel extracted in the feature amount extraction step, the corresponding step of associating the vehicle in which the wheel is imaged in the first point image with the vehicle in which the wheel is imaged in the second point image. And let the computer run
In the feature amount extraction step, resizing correction is performed to convert the wheels imaged in the first point image and the second point image into a reference size, and the resizing-corrected image of the wheels is subjected to resizing correction. An identification program that is a step of extracting the feature amount of the wheel as the feature amount.

Images