JP7302825B2 - Combined data determination system and combined data determination program - Google Patents

Description

The present invention relates to combinatorial data systems, and more particularly to combinatorial estimation processing.

In Patent Document 1, WCN (wireless call number) data of the on-board device of the vehicle and automobile registration number of the vehicle are obtained by communicating with the ETC on-board device mounted on the vehicle by the DSRC (Dedicated Short Range Communications) system. (hereinafter referred to as "vehicle number").

However, in reality, it is not possible to obtain the combined data of the WCN data and the vehicle number through communication with the ETC vehicle-mounted device.

Therefore, the inventor has come up with the idea of capturing the vehicle number of a vehicle traveling on the road and communicating with the ETC on-board device of the vehicle to acquire the combination data.

JP-A-2009-37468

However, it turned out that there exist the following problems about acquisition of the said combination data.

In the first place, the analysis of the vehicle number from the imaged image data may fail. Also, WCN data may not be received. Furthermore, with respect to the combination data, the WCN data received with a certain time difference is associated with the time when the vehicle number was captured. Therefore, when a plurality of vehicles pass together, there is a possibility that the photographed body number and WCN data cannot be associated one-to-one.

Such a problem becomes more pronounced when common equipment is installed in each lane.

SUMMARY OF THE INVENTION It is an object of the present invention to solve the above problems and to provide a combination data determination system capable of reliably obtaining combination data of vehicle number and WCN data.

(1) A combination data determination system according to the present invention uses a plurality of WCN data acquired from a vehicle by vehicle information acquisition means as WCN group data, and a plurality of vehicle number data detected in association with the WCN group data. Candidate group data storage means for storing, as number group data, a plurality of candidate group data having different detection times or detection locations for the candidate group data composed of the WCN group data and the vehicle number group data ;
Candidate combination data generation means for generating a plurality of combination data of WCN data of the WCN group data and vehicle number data of the vehicle number group data as candidate combination data for each of the candidate group data. ,
When it is assumed that certain candidate combination data among the candidate combination data is correct, if there is no contradiction in the relationship with other candidate combination data in the candidate group data including the assumed candidate combination data, the assumed candidate combination data A combination determining means for determining candidate combination data is provided.

Therefore, even if a plurality of combinations of combination data of vehicle number and WCN data are detected, they can be reliably associated with each other.

(2) In the combination data determination system according to the present invention, the combination determination means has the same data in either the vehicle number or the WCN data constituting the focused candidate combination data in the plurality of candidate group data. Other candidate group data are extracted, and if all of the extracted candidate group data have the focused candidate combination data, the candidate combination data are determined. Thus, consistent associations of detected results at different times or locations are possible.

(3) In the combination data determination system according to the present invention, the combination determination means removes the determined candidate combination data and repeats the determination process. By removing newly established candidate combination data, the likelihood of combinations can be reduced.

(4) In the combination data determination system according to the present invention, the combination determination means, when assuming that certain candidate combination data is correct, determines whether the candidate group data including the hypothesized candidate combination data is combined with other combination data . If there are a plurality of combination data that do not cause contradiction in relation, such candidate combination data are removed, and the process of determining the assumed candidate combination data is repeated. Therefore, the possibility of combinations can be reduced.

(5) In the combination data determination system according to the present invention, the combination determination means makes the determination when there is no missing candidate combination data in the candidate group data. Therefore, it is possible to prevent erroneous association when there is a defect.

(6) In the combination data determination system according to the present invention, the combination determination means selects candidates for the case where there is no missing candidate combination data in the candidate group data, and where there are a plurality of combinations that do not cause contradiction. Remove combination data. Therefore, it is possible to prevent erroneous association when there is a defect.

(7) In the combination data determination system according to the present invention , if there is a contradiction with already determined candidate combination data, it is stored as data to be replaced. Therefore, when the correspondence between the vehicle number and the WCN data is changed, it can be subject to washing.

(8) A combination data determination program according to the present invention is a combination data determination program for configuring a computer as the following means. A plurality of WCN data obtained from a vehicle by a vehicle information obtaining means are defined as WCN group data, and a plurality of vehicle number data detected in association with the WCN group data are defined as vehicle number group data, and the WCN group data and the vehicle number are defined as vehicle number group data. Candidate group data storage means for storing a plurality of candidate group data different in detection time or detection location for candidate group data composed of group data; and for each candidate group data, one WCN data of the WCN group data candidate combination data generating means for generating a plurality of combination data of one vehicle number data from the vehicle number group data as candidate combination data; (2) combination determination means for determining the assumed candidate combination data when there is no contradiction in relation to other candidate combination data in the candidate group data containing the assumed candidate combination data;

Therefore, even if a plurality of combinations of combination data of vehicle number and WCN data are detected, they can be reliably associated with each other.

(9) In the combination data determination system according to the present invention, when the candidate group data includes missing candidate combination data, the combination determination means selects candidate combination data that is not missing from the candidate group data. is extracted, and other candidate group data having the same data as either the vehicle number or the WCN data constituting the focused candidate combination data is extracted, and all of the extracted candidate group data are combined with the focused candidate combination data and a point based on the number of the extracted candidate group data is obtained, and if the number exceeds the threshold value, the candidate combination data is determined.
Therefore, the combination data can be determined from candidate group data having missing candidate combination data.

(10) In the combination data determination system according to the present invention, when the extracted candidate group data differ in detection date or detection location, adjustment is made so that the points are higher. Therefore, if the detection dates or detection locations are different, the combination data can be determined from a small number of candidate group data.

(11) In the combination data determination system according to the present invention, if the number of missing candidate combination data included in the candidate group data is less than a predetermined number, the points are adjusted to be higher. Therefore, for candidate group data with few omissions, the combination data can be determined from a small number of candidate group data.

(12) In the combination data determination system according to the present invention, the combination determination means may be configured such that, for the plurality of candidate group data, either the vehicle number or the WCN data constituting the focused candidate combination data has the same data. of the candidate group data as first candidate group data, the number thereof is calculated, and among the first candidate group data, candidate group data containing the same candidate combination data as the noted candidate combination data is selected as a second candidate group data The number of candidate group data is obtained, the evaluation judgment value is calculated based on the value obtained by dividing the number of the second candidate group data by the number of the first candidate group data, and the evaluation judgment value is designed in advance. If the calculated threshold value is exceeded, the candidate combination data is determined.
Therefore, even if the combination of the vehicle number or the WCN data is wrong due to the detection time lag, it is possible to prevent the wrong candidate combination data from being determined.

(13) In the combination data determination system according to the present invention, the combination determination means determines the second candidate whose detection date or detection location is different when calculating the evaluation judgment value based on the number of the second candidate group data. If there is group data, the evaluation judgment value is adjusted to be higher. Therefore, even if the combination of the vehicle number or the WCN data is wrong due to the detection time lag, it is possible to more reliably prevent the determination of wrong candidate combination data.

(14) In the combination data determination system according to the present invention, the combination determination means unites the candidate group data at the same point at a similar time to form one candidate group data. Therefore, at least correct candidate combination data can be generated by being treated as one candidate group data.

(15) In the combination data determination program according to the present invention, the combination determination means determines that either the vehicle number or the WCN data constituting the candidate combination data of interest has the same data for the plurality of candidate group data. of the candidate group data as first candidate group data, the number thereof is calculated, and among the first candidate group data, candidate group data containing the same candidate combination data as the noted candidate combination data is selected as a second candidate group data The number of candidate group data is obtained, the evaluation judgment value is calculated based on the value obtained by dividing the number of the second candidate group data by the number of the first candidate group data, and the evaluation judgment value is designed in advance. If the calculated threshold value is exceeded, the candidate combination data is determined.
Therefore, even if the combination of the vehicle number or the WCN data is wrong due to the detection time lag, it is possible to prevent the wrong candidate combination data from being determined.

(16) In the combination data determination program according to the present invention, when the combination determination means obtains the evaluation judgment value based on the number of the second candidate group data, the second candidates whose detection date or detection location is different. If there is group data, the evaluation judgment value is adjusted to be higher. Therefore, even if the combination of the vehicle number or the WCN data is wrong due to the detection time lag, it is possible to more reliably prevent the determination of wrong candidate combination data.

(17) In the combination data determination program according to the present invention, the combination determination means unites the candidate group data at the same site at close times to form one candidate group data. Therefore, at least correct candidate combination data can be generated by being treated as one candidate group data.

The meaning of the terms used in this specification will be explained. “Acquired in association with” means acquired within a certain width of detection time.

"Candidate combination data" refers to candidates that can be combined from the vehicle number and WCN data acquired at the same time, and in the embodiment, exist for each candidate group data in FIG.

"Candidate group data" is data composed of a plurality of vehicle numbers and a plurality of WCN data, and corresponds to candidate group data K1 to K7 in FIG. 6A in this embodiment.

“Candidate combination data with missing” refers to candidate combination data in which one of the data described in FIGS. 6 and 15 is “missing” in the embodiment.

"Evaluation judgment value" refers to the score calculated in step S305 of FIG. 21 in the embodiment, and is a concept that includes both the occurrence rate and the total number of points.

1 is a diagram showing the overall configuration of a dispatched vehicle determination system 1; FIG. 2 is a functional block diagram of an in-vehicle device 200 and a roadside device 300 of the dispatched vehicle determination system 1; FIG. 3 is a diagram showing an example of a hardware configuration in which the management server 100 is implemented using a CPU; FIG. It is a figure which shows the data structure of detection information data. 8 is a flowchart of detection information storage processing; It is an example of combination data of WCN data and a vehicle number. 1 is a functional block diagram of a combination data determination system 1; FIG. It is a flow chart of combination confirmation processing. It is a condition 1 matching processing flowchart. It is an example of candidate combination data for each candidate group data. It is an example of partially deleted candidate combination data. It is an example of partially deleted candidate combination data. It is a condition 2 matching processing flowchart. It is a condition 3 agreement process flowchart. It is an example of combination data of WCN data and a vehicle number. It is an example of combination data of WCN data and a vehicle number in 2nd Embodiment. It is an example of combination data of WCN data and a vehicle number in 2nd Embodiment. 10 is a flowchart of combination confirmation processing in the second embodiment; It is an example of candidate combination data for each candidate group data. It is an example of combination data of WCN data and a vehicle number in 2nd Embodiment. It is a flow chart of score judgment processing.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described with reference to the drawings.

A. Description of Prerequisite Vehicle Determination System A brief description will be given of the presupposed vehicle determination system of the present invention.

(A1. Overall structure)
FIG. 1 shows the configuration of a dispatched vehicle determination system 1. As shown in FIG.

The dispatched vehicle determination system 1 is composed of a management server 100 and a plurality of roadside devices 300 arranged at different positions. Each road device 300 acquires vehicle identification information including at least WCN data acquired from vehicles located within an acquisition range and vehicle number data obtained by imaging the vehicle number. The roadside device 300 is connected to the management server 100 via a network, and transmits the acquired WCN data and the vehicle number of the vehicle to the management server 100 .

As a result, it is possible to acquire combination data of the WCN data and the vehicle number of the vehicle.

(A2. Acquisition of WCN data and vehicle number)
Acquisition of WCN data and vehicle number will be described with reference to FIG. FIG. 2A shows a block diagram of an in-vehicle device 200 installed in each vehicle. The in-vehicle device 200 is an ETC (Electronic Toll Collection System) in-vehicle device and has a DSRC communication function. The in-vehicle device 200 includes a control section 202 , a wireless communication section 203 , a card reading section 204 and a storage section 206 .

The control unit 202 has a CPU and memory, and controls each unit of the in-vehicle device 200 . The wireless communication unit 203 is a communication unit that enables wireless communication by DSRC, and has an antenna 203a for wireless communication. The control unit 202 stores WCN data, which is identification information of its own device.

A block diagram of the roadside device 300 is shown in FIG. 2B. The roadside device 300 has a DSRC communication function, a wired communication function, and an imaging function. The roadside device 300 includes a control section 301 , a wireless communication section 302 , a wired communication section 303 , a storage section 304 and an imaging section 305 .

The control unit 301 is configured by a CPU, memory, etc., and controls each unit of the roadside device 300 .

The wireless communication unit 302 is a communication unit that enables wireless communication with a vehicle that has entered the wireless communication area 220 (see FIG. 1) by DSRC, and has an antenna 302a for wireless communication. The wired communication unit 303 communicates with the connected management server 100 via the communication network.

The imaging unit 305 includes a camera (not shown) and captures a moving image of the vehicle entering the wireless communication area 220 shown in FIG. The control unit 301 analyzes the video data and acquires the vehicle number if it exists. Also, the imaged vehicle transmits paired data with the WCN data received at the time when it entered the wireless communication area 220 to the management server 100 via the wired communication unit 303 .

The storage unit 304 stores information such as video data received from the in-vehicle device 200 .

In this embodiment, moving image data is analyzed, but continuous still images may be picked up and the vehicle number extracted from the still images.

(A3. Hardware configuration)
A hardware configuration of the management server 100 shown in FIG. 1 will be described using FIG. The figure shows an example of a hardware configuration in which the management server 100 is configured using a CPU.

The management server 100 has a CPU 23 , a memory 27 , a hard disk 26 , a monitor 30 , a communication board 31 , an input device 28 , an optical drive 25 and a bus line 29 . The CPU 23 controls each part via the bus line 29 according to each program stored in the hard disk 26 .

The hard disk 26 stores an operating system program 26o (hereinafter abbreviated as OS) and a main program 26p. The hard disk 26 has a detection information database 26s, a detection base information database 26g, and a storage section 26k.

Processing of the main program 26p will be described later.

FIG. 4A shows the data structure of detection information data stored in the detection information database 26s. In this embodiment, the detection information data is composed of the items "detection ID", "detection time", "detection site ID", "antenna number", "vehicle number" and "WCN data".

"Detection ID" is a unique identifier given in order of detection. For each detection ID, items "detection time", "detection base ID", and "antenna number" are stored together with the vehicle identification information received from each roadside device 300. FIG. The reason why one "detection base ID" has multiple "antenna numbers" is that one roadside device may be provided with multiple pairs of cameras and wireless communication units that perform wireless communication using DSRC. . The item “detection base ID” is a unique ID for each roadside device 300 .

In FIG. 4A, there are cases where one WCN data is associated with one vehicle number, and a plurality of WCN data are associated with a plurality of vehicle numbers. These will be described later.

Position information of the roadside device 300 is stored in the detection base information database 26g. Specifically, it consists of a "detection base ID" and "location information".

Although LINUX (registered trademark or trademark) is used as the operating system program (OS) 26o in this embodiment, it is not limited to this.

Each of the above programs is read from an optical disk (CD-ROM 25a or the like) storing the programs via the optical drive 25 and installed in the hard disk 26. FIG. In addition to the CD-ROM, a program such as a flexible disk (FD) or an IC card may be installed from a computer-readable recording medium to a hard disk. Furthermore, it may be downloaded using a communication line.

In this embodiment, by installing the program from the optical disk to the hard disk 26, the computer is indirectly caused to execute the program stored in the optical disk. However, without being limited to this, a program stored on an optical disk may be directly executed from the optical drive 25 . In addition, programs that can be executed by a computer include those that can be directly executed by simply installing them as they are, as well as those that need to be converted into other formats once (for example, decompressing compressed data). ), and also those that can be executed in combination with other module parts.

(A4. Detection information storage processing)
The processing for generating the detection information shown in FIG. 4A will be described with reference to FIG.

The control unit 301 shown in FIG. 2B determines whether or not to detect a vehicle in the imaging data given from the imaging unit 305 (step S1 in FIG. 5). When a vehicle is detected in the imaging data, the vehicle number is analyzed and stored (step S3). Any method can be used to analyze the vehicle number from the image data, and known means can be employed. FIG. 4C shows the data structure of the stored data. In the present embodiment, combination data of imaging time and vehicle number is stored. If the vehicle is not detected in step S1, the process proceeds to step S5.

Control unit 301 determines whether or not WCN data has been detected (step S5). In this embodiment, when a vehicle enters the wireless communication area 220 shown in FIG. 1, it is determined that the WCN data has been detected. When the WCN data is detected, the control unit 301 stores the WCN data together with the detection time (step S6). FIG. 4B shows the data structure of the stored data. In this embodiment, combination data of detection time and WCN data is stored.

If WCN data is not detected in step S5, the process proceeds to step S1, and steps S1 and subsequent steps are repeated.

The control unit 301 determines whether or not there is a matching combination (step S9). In this embodiment, the matching combination is a combination of WCN data detected at the same time as the vehicle number imaging time. It should be noted that the determination as to whether or not the time is the same may be made with some leeway.

Moreover, as such a matching combination, a method other than the matching of the times may be used. For example, a vehicle may be equipped with a GPS device to transmit location information together with WCN data to the roadside device 300, so that a combination with matching location information may be selected.

If there is a matching combination in step S7, the control unit 301 creates detection information by adding "detection point ID", "antenna number", and "antenna direction" to the combination, and transmits it to the server ( step S9).

On the other hand, if there is no matching combination in step S7, the control unit 301 transmits detection information data with one missing to the management server 100 (step S11).

Upon receiving such detection information data, the management server 100 assigns a detection ID and adds new detection information data to the detection information database shown in FIG.

Thereby, the management server 100 can acquire combined data of the vehicle number detected by the roadside device 300 and the WCN data.

For example, in FIG. 4, as the combined data of the vehicle number and WCN data, there are cases where a plurality of WCN data are associated with a plurality of vehicle numbers. For example, the detection ID "1002" has four vehicle numbers "Kobe 300 or 3124", "Kobe 500 1026", "Osaka 320 5467", "Izumi 333 1025", and four WCN data "876543210001", "942123545601", "689456289423", and "478541236912" are associated data.

In this embodiment, the WCN data is obtained after imaging, but the order of obtaining these data is arbitrary.

Also, the case where the control unit 301 of the roadside device 300 creates combination data of the vehicle number and the WCN data has been described, but the control unit 301 includes the WCN data specifying the detection time shown in FIGS. The detection result of the vehicle number specifying the time may be transmitted to the management server 100, and the management server 100 may make such correspondence.

In addition, in FIG. 5, there are cases where the vehicle body number cannot be found even after the analysis at S3. Also, in some cases, detection of WCN data fails in S5. In that case, it may be transmitted to the management server 100 as combined data in which one of them is missing.

As described above, in the above embodiment, the vehicle number is associated with the WCN data received with a certain time difference from the time when the vehicle number was captured. In addition, the roadside device 300 acquires the WCN data and vehicle number of each vehicle, and is shared by multiple lanes. Therefore, with regard to the combined data of the WCN data and the vehicle number, there is a possibility that multiple-to-multiple data detection is performed for the vehicle number and the WCN data as shown in FIG. 6A. In the following, a process of specifying an identifiable combination from such a plurality of combination candidates is performed.

B. First Embodiment (B1. Functional Block Diagram)
An overview of the combined data determination system 1 will be described with reference to FIG. The combination data determination system 1 determines a combination of the WCN data and vehicle number acquired in association by the vehicle information acquisition means 300 . The combination data determination system 1 includes candidate group data storage means 4 , candidate combination data generation means 5 and combination determination means 6 .

Candidate group data storage means 4 stores candidate group data composed of a plurality of WCN data acquired from vehicles by vehicle information acquisition means 300 and a plurality of vehicle number data detected in association with the WCN data at detection time. Alternatively, a plurality of candidate group data with different detection locations are stored. The candidate combination data generation means 5 generates a plurality of combination data of one WCN data and one vehicle number data as candidate combination data for each of the candidate group data. Combination determining means 6 Determines the assumed candidate combination data when it is assumed that certain candidate combination data among the candidate combination data is correct and there is no contradiction in relation to other combinations in the candidate group data. Let Specifically, from the plurality of candidate group data, other candidate group data having the same vehicle number or WCN data constituting the focused candidate combination data is extracted, and extracted candidate group data have the noted candidate combination data, the candidate combination data may be confirmed, the confirmed candidate combination data may be removed, and the confirmation process may be repeated. .

Further, when it is assumed that certain candidate combination data is correct, if there are a plurality of combinations that do not cause contradiction in relation to other combinations in the candidate group data, it is semi-definite within these ranges, and The confirmation process may be repeated by removing semi-definite candidate combination data.

The combination determining means 6 may perform the confirmation or semi-decision when there is no missing candidate combination data in the candidate group data.

In addition, the combination determining means 6 stores the candidate combination data, which has already been confirmed, as an object to be replaced if it contradicts the already established candidate combination data.

In this way, by performing the confirmation process with reference to candidate group data with different detection times or detection locations, WCN data and vehicle numbers can be identified from a combination of a plurality of group data.

(B2. Combination Confirmation Process)
The association of the combination data by the management server 100 will be described. The main program 26p of the management server 100 includes a program for performing the combination confirmation process shown in FIG. 8, and the CPU 23 performs the detection information data generation process in step S7 of FIG. 5 based on the association process program. Here, a case where a plurality of candidate combination data as shown in FIG. 6A is detected will be described as an example. In the following, to simplify the explanation, the vehicle number and WCN data are abstracted as N○ and W○, respectively.

The candidate combination data shown in FIG. 6A will be briefly described.

Since the candidate group data K1 is (N1, N2, N3, N4, N5)*(W1, W2, W3, W4, W5), the following 25 candidate combination data (N1, W1), (N1, W2 ), (N1, W3), (N1, W4), (N1, W5), (N2, W1), (N2, W2), (N2, W3), (N2, W4), (N2, W5), (N3, W1), (N3, W2), (N3, W3), (N3, W4), (N3, W5), (N4, W1), (N4, W2), (N4, W3), (N4 , W4), (N4, W5), (N5, W1), (N5, W2), (N5, W3), (N5, W4), (N5, W5).

Similarly, candidate data group candidate group data K2 is (N6, N7, N8, N9, N10)*(W6, W7, W8, W9, missing 1), so the following 25 candidate combination data (N6, W6), (N6, W7), (N6, W8), (N6, W9), (N6, deficit 1) ... (N10, deficit 1) can be established.

Also, since the candidate group data K6 is (N1, N20)*(W1, W21), the following four candidate combination data (N1, W1), (N1, W21), (N20, W1), (N20 , W21) can hold.

Also, since the candidate group data K7 is (N16, missing 3, N18, N19)*(W14, W16, W18, W19), the following 16 candidate combination data (N16, W14), (N16, W16) , (N16, W18), (N16, W19), (missing 3, W14), (missing 3, W16), (N19, W19).

Here, "missing" means that the WCN data cannot be received in the first place because the target vehicle is not equipped with an ETC on-board device, or the WCN data could not be received for some reason. . Regarding the vehicle number, there are cases where the image could be captured but could not be analyzed. In addition, cases where such defects occur are not limited to these.

FIG. 6B shows existing definite data associated one-to-one with the detections so far. Here, it is assumed that there are three existing fixed data P1 to P3.

The combination confirmation process will be described with reference to FIG.

The CPU 23 reads each candidate group data and creates candidate combination data for each candidate group data (step S103). In this case, candidate combination data described above is created for all candidate group data K1 to K7.

The CPU 23 deletes existing determined data for the created candidate combination data, if any (step S105). In this case, the existing fixed data are P1 (N19, W19), P2 (N18, W18), and P3 (N20, W20), so these and the combined data whose possibility is eliminated by fixing these are removed. be. For example, in this case, the candidate combination data (N19, W19) and (N18, W18) exist in the candidate group data K7. Candidate combination data (N19,W14), (N19,W16), (N16,W18), (N18,W14), (N18) containing any of N19), (W19), (N18), (W18) , W16) and (N18, W19) are also deleted.

The CPU 23 sets the attention group data number i=0 (step S107), and reads out all the candidate combination data of the attention group data (step S109).

The CPU 23 determines whether or not the candidate combination data of the attention group data is 1 combination data (step S111). In this case, since the 0th attention group data K1 is 25 combination data, the CPU 23 sets the attention candidate number j=0 (step S113).

The CPU 23 reads the j-th candidate combination data (step S115). In this case, the 0th candidate combination data (N1, W1) is read. The CPU 23 determines whether the candidate combination data (N1, W1) meets condition 1 (step S117). FIG. 9 shows the details of the matching determination of condition 1. FIG.

The CPU 23 extracts candidate group data having the same vehicle number N1 as the candidate combination data (N1, W1) (step S201). In this case, since the same vehicle number N1 exists in candidate group data K1, K3, K4, K5, and K6, candidate combination data of these candidate group data is extracted. The CPU 23 deletes candidate group data including missing data from the extracted candidate group data (step S203). In this case, since candidate group data K3 contains a defect, it is deleted from the extraction target, and candidate combination data of candidate group data K1, K4, K5, and K6 remains as shown in FIG.

The CPU 23 determines whether or not there is candidate combination data common to the extracted candidate group data (step S205). In this case, the candidate combination data (N1, W1) exists in all four candidate group data K1, K4, K5, K6, so the CPU 23 determines whether there is only one such candidate (step S207). ). In this case, other than such candidate combination data, there is no candidate combination data common to all candidate group data .

The CPU 23 determines whether the candidate combination data (N1, W1) is consistent with the existing fixed data (step S209). In this case, the existing fixed data are combination data (N19, W19), (N18, W18), (N20, W20), so there is no contradiction. Therefore, the CPU 23 determines the candidate combination data (N1, W1) (step S211).

The CPU 23 deletes the candidate combination data including either the vehicle number N1 or the WCN data W1 forming the candidate combination data (N1, W1) from all the candidate group data K1 to K7 (step S213). Further, the candidate combination data including either the vehicle number N1 or the WCN data W1 constituting this combination data is deleted from the candidate combination data of all the candidate group data K1 to K7.

FIG. 11 shows candidate combination data after deletion of candidate group data K1, K4, K5, and K6 shown in FIG. As a result, the process of step S117 in FIG. 8 ends.

The CPU 23 determines whether or not Condition 1 is met (step S119). In this case, since condition 1 is met, the CPU 23 determines whether or not the target candidate number j is the last (step S131).

In this case, since the attention candidate number j is not the last, the CPU 23 increments the attention candidate number j (step S133), and reads the candidate combination data of j=1 (step S115). In this case, candidate combination data (N2, W2) next to the candidate combination data remaining in step S213 of FIG. 9 is read out, and it is determined that condition 1 is satisfied in the same manner as above (step S117).

In this case, candidate group data K1, K4 having vehicle number N2, which is an element of candidate combination data (N2, W2), is extracted (step S201 in FIG. 9).

In this case, the candidate group data K1 and K4 are candidate combination data as shown in FIG. 12A.

Among the extracted candidate group data , the CPU 23 deletes the candidate group data including the defect from the extraction target (step S203). In this case, since there is no candidate group data including a defect, the CPU 23 determines whether or not there is candidate combination data common to the extracted candidate group data (step S205). In this case, the candidate combination data (N2, W2) exists in both of the two candidate group data K1, K4. Therefore, the CPU 23 determines whether or not one such candidate can be confirmed (step S207).

In this case, candidate combination data (N2, W3), (N3, W2), and (N3, W3) exist as common candidate combination data in addition to candidate combination data (N2, W2). , cannot be fixed to one, the condition 1 match determination process is terminated.

The CPU 23 determines whether or not Condition 1 is met (step S119 in FIG. 8). In this case, since the condition 1 is not met, the CPU 23 performs condition 2 match determination processing (step S121).

The CPU 23 extracts candidate group data having the same vehicle number N2 from the candidate combination data (N2, W2) (step S251). In this case, candidate group data K1 and K4 are extracted. The candidate group data K1 and K4 are candidate combination data as shown in FIG. 12A. The CPU 23 deletes the candidate group data including the loss from the extracted candidate group data (step S253). In this case, nothing is deleted because there is no candidate group data containing a defect.

The CPU 23 determines whether or not there is candidate combination data common to the extracted candidate group data (step S255). In this case, candidate combination data (N2, W2) exists in both candidate group data K1 and K4.

The CPU 23 determines whether or not there are N:N such candidates (step S257). In this case, there are candidate combination data (N2, W3), (N3, W2), (N3, W3). Since the number of vehicle numbers and the number of WCN data in the candidate combination are the same (two N and two W are the same number), the CPU 23 puts these candidate combination data in a semi-definite state (step S259), and delete from the target (step S261).

Further, as in step S213 in FIG. 9, all candidate combination data including either the element (N1) or the element (W1) constituting this combination data are deleted from all the candidate combination data. As a result, candidate combination data as shown in FIG. 12B remain in the candidate group data K1 and K4.

In this way, even if one cannot be determined, it can be deleted from the candidate group data by semi-determining it, so that the remaining candidate combination data can be reduced.

If the same number of N:N combinations do not exist in step S257 of FIG. 13, the process is terminated.

The CPU 23 returns to FIG. 8 and determines whether or not Condition 2 is met (step S121 in FIG. 8). In this case, since condition 2 is met, the CPU 23 determines whether or not the target candidate number j is the last (step S131).

In this case, since the attention candidate number j is not the last, the CPU 23 increments the attention candidate number j (step S133), and reads the candidate combination data of j=2 (step S115). In this case, candidate combination data (N4, W4) next to the candidate combination data remaining in step S215 in FIG. 9 is read out, and it is determined that condition 1 is satisfied in the same manner as above (step S117). In this case, the candidate group data K1, K3 having the vehicle number N4 of the candidate combination data (N4, W4) are read out (step S201 in FIG. 9). Among the extracted candidate group data , the CPU 23 deletes the candidate group data including the defect from the extraction target (step S203). In this case, since the candidate group data K3 contains a defect, it is deleted from the extraction target, leaving the candidate combination data of the candidate group data K1.

The CPU 23 determines whether or not there is candidate combination data common to the extracted candidate group data (step S205). In this case, since there is only one candidate group data, there is no common candidate combination data. Therefore, the CPU 23 terminates the matching determination of Condition 1. FIG.

The CPU 23 determines whether or not the condition 1 is met (step S119 in FIG. 8), and since it is not met in this case, it determines whether the condition 2 is met (step S121). As with condition 1, condition 2 also terminates the process at step S255 in FIG.

The CPU 23 returns to FIG. 8 and determines whether or not Condition 2 is met (step S123 in FIG. 8). In this case, since the condition 2 is not satisfied, the CPU 23 determines the condition 3 (step S125).

The matching determination of condition 3 will be described with reference to FIG. 14 .

The CPU 23 extracts other candidate group data in which there is a combination having the same vehicle number as the vehicle number N4 of the candidate combination data (N4, W4) (step S281). In this case, candidate group data K1 and K3 are read.

The CPU 23 determines whether or not the candidate group data K1 and K3 that have been read include candidate combination data having a defect (step S283). In this case, since the candidate group data K3 includes candidate combination data having a defect, the CPU 23 extracts candidate group data having other candidate combination data included in the candidate group data K3 (step S285). In this case, candidate group data K3 includes candidate combination data (N4, W11), (N4, W12), (N4, missing 2), (N11, W11), (N11) in addition to candidate combination data (N4, W4). , W12), (N11, deficiency 2), (N12, W11), (N12, W12), (N12, deficiency 2), (N13, W11), (N13, W12), (N1, deficiency 2) . Therefore, candidate group data K5 having candidate combination data (N11, W11) is extracted from these.

The CPU 23 determines whether or not a single matching is possible from the candidate combination data of these three candidate group data and the already semi-confirmed candidate combination data (step S287). In this case, condition 3 is terminated because a single match cannot be made.

The CPU 23 determines whether or not the target candidate number j is the last (step S131 in FIG. 8).

In this case, since the attention candidate number j is not the last, the CPU 23 increments the attention candidate number j (step S133), and reads the candidate combination data of j=3 (step S115). In this case, the next candidate combination data (N4, W5) is read, and it is determined whether the conditions 1 to 3 are met in the same manner as described above.

(N5, W5) are read in the same manner, and when the matching judgment of the conditions 1 to 3 is completed in the same manner as above, the CPU 23 judges whether or not all the candidate group data have been examined (step S141).

In this case, since all the candidate group data have not been examined, the CPU 23 increments the attention group data number i and repeats the processing from step S109.

Since the examination of the candidate combination data (N16, W11) of the candidate group data K5 is the same as the above, the explanation is omitted.

A case where candidate combination data (N16, W16) is a target candidate will be described. In this case, conditions 1 and 2 shown in FIG. 9 are not satisfied. This is because the candidate group data K7 having the same candidate combination data (N16, W16) contains a defect.

Therefore, the CPU 23 determines condition 3 (step S125 in FIG. 8). The CPU 23 extracts other candidate group data in which there is a combination having the same vehicle number as the vehicle number N16 of the candidate combination data (N16, W16) (step S281). As a result, candidate group data K5 and K7 are read.

The CPU 23 determines whether or not the read candidate group data K5 and K7 include candidate combination data having a defect (step S283). In this case, since the candidate group data K7 includes the missing candidate combination data, the CPU 23 extracts candidate group data having other candidate combination data included in the candidate group data K5 and K7 (step S285).

In this case, candidate group data K7 has candidate combination data (N16, W14), (loss 3, N14), (loss 3, N16) in addition to candidate combination data (N16, W16). Therefore, the candidate group data K4 having the vehicle number N14 or the WCN data W14) constituting the candidate combination data is extracted from these.

The CPU 23 determines whether or not a single matching is possible from the candidate combination data of these three candidate group data and the already semi-confirmed candidate combination data (step S287).

Candidate group data K5 has candidate combination data (N1, W1) deleted, and has candidate combination data (N6, W6), (N6, W11), (N6, W16), (N6, W17), (N11, W6), (N11, W11), (N11, W16), (N11, W17), (N16, W6), (N16, W11), (N16, W16), (N16, W17), (N17, W6) , (N17, W11), (N17, W16), (N17, W17) remain.

On the other hand, in candidate group data K7, candidate combination data (N18, W18) and (N19, W19) are deleted, and candidate combination data (N16, W14), (N16, W16), (missing 3, W14), ( Defect, W16) remains.

In addition, candidate combination data (N14, W14), (N14, W15), (N15, W14), and (N15, W15) remain in candidate group data K4.

Based on the candidate group data K4, the CPU 23 determines that the candidate combination data (N16, W14) does not exist, confirms the candidate combination data (N16, W16), and deletes it from the target (step S289).

The CPU 23 determines whether or not the remaining single matches are possible (step S291). In this case, the candidate combination data (deficit 3, W14) can be matched with a single match, so it is determined and deleted from the target (step S293).

In this way, a single matching result can be obtained even for candidate combination data in which one of the vehicle number and WCN data is missing.

When all the candidate group data have been examined, the conditions 1 to 3 are again determined for the remaining candidate group data. This is because even if the finalization process cannot be performed at a certain timing, it may be finalized when another candidate combination data is finalized, because the number of targets decreases.

The CPU 23 repeats the process of step S144 as long as there are items that can be additionally confirmed. When there is nothing that can be fixed, the process ends.

A brief description will be given of the processing in step S209 in FIG. For example, in the candidate group data K6, when the candidate combination data (N1, W1) is determined, the candidate combination data (N20, W21) can be single-matched. However, this combination contradicts existing established combination data (N20, W20). Such situations may arise, for example, if the equipment that emits the WCN data is changed, if the vehicle changes ownership and is given a different vehicle number, or if it has a counterfeit license plate. It is from. In this embodiment, if there is a possibility of such a change, it is stored as a replacement candidate (step S217 in FIG. 9).

If only one set of combination data is detected at step S111 in FIG. 8, the detection information is generated as it is and the process ends (step S149).

It is also possible to specify conditions other than those adopted in this embodiment. For example, in condition 3, the processing of steps S285 to S293 may be performed only when a loss is included, or the processing of steps S285 to S293 may be performed even when no loss is included.

(B3. Processing when there is a defect)
In the above-described embodiment, candidate group data with no defects are used, and if they can be determined, they are determined as combined data. This is because if there are only two pieces of candidate group data in which a defect exists (hereinafter referred to as "missing candidate group data"), it cannot be determined that the combination is reliable. However, even in such defect-containing candidate group data, if such a combination exists in many defect-containing candidate group data, it may be determined as one combination. For example, when there is an identical combination of defect-containing candidate group data, the point is +1, and the total exceeds the threshold value of 10.

In this case, if the same combination data exists, points may be added and it may be determined whether the number exceeds a predetermined threshold.

(B4. Weighting process)
When judging combination data based on the above points, weighting conditions are stored according to whether the number of defects in the candidate group data containing defects and the date or place of detection of the candidate group data containing defects are the same. A weight may be assigned for determination.

Weighting rules can be set as follows, for example.

If the data was detected on a different day, add the number of days detected x 0.5.

・If the data is detected at another site, add the total number of detected sites x 0.5.

- +0.5 for candidate group data with 1 defect, no addition for candidate group data with 2 to 4 defects, -0 for candidate group data with 5 or more defects .5.

In the case of FIG. 15, when the candidate combination data (N30, W30) of the defect-containing candidate group data K11 is being examined in the defect-containing candidate group data K11 to K15, in the defect-containing candidate group data K12 to K15, The candidate combination data (N30, W30) exists.

Here, the defect-containing candidate group data K12 is the defect-containing candidate group data having one defect on the same day and at the same site as the defect-containing candidate group data K11. Therefore, the point becomes +1.5. The defect-containing candidate group data K13 differs from the defect-containing candidate group data K11 in detection date and detection location, but since it includes two defects, the points are 1.5+1.5=3. The defect-containing candidate group data K14 differs from the defect-containing candidate group data K11 in detection date and detection location, and has one defect, so 1.5+1.5+0.5=3.5. The defect-containing candidate group data K15 differs from the defect-containing candidate group data K11 in detection date and detection location, but since it includes two defects, the points are 1.5+1.5=3.

This gives a total of 1.5+3+3.5+3=11 points. Therefore, since the set threshold 10 is exceeded, the candidate combination data (N30, W30) can be determined.

By setting weighting rules in this way, it is possible to determine combination data with less defect-containing candidate group data.

In the above embodiment, points are added on different days, but points may be added on the same day after a certain amount of time has passed.

The above degree of point addition is an example, and can be changed as appropriate. For example, if the distance between bases is long, the points may be increased. For example, if the distance between bases is only about 300 meters away, there is no addition, and if the distance is more than 1 km, the distance is added.

In the above-described embodiment, combination data is determined using only candidate group data that does not include a defect, and then determination is made based on defect-containing candidate group data. The order of the processing to be performed is arbitrary. For example, it may be performed before step S117 in FIG.

(B5. Other embodiments)
In the present embodiment, the case where about 25 sets of candidate combination data (5*5) are generated in the candidate group data has been described as an example. If it is assumed that certain candidate combination data is correct, instead of the judgment as in condition 1, if there is no contradiction in relation to other combinations in the candidate group data , the assumed candidate combination data is determined. can be

For example:

If the candidate group data K20 is (N201, N202)*(W201, W202) and the candidate group data K30 is (N201, N230)*(W201, W230), the former is (N201 ,W201), (N201,W202),(N202,W201), (N202,W202) and the latter is (N201,W201), (N201,W230),(N230,W201), (N230,W230) . In this case, assuming that candidate combination data (N201, W201) is correct, no contradiction occurs in other candidate group data K30. On the other hand, if candidate combination data (N201, W202) is correct, contradiction occurs in other candidate group data K30.

In the present embodiment, in conditions 1 to 3, the vehicle number is focused and the candidate group data having the same vehicle number are extracted (for example, step S201 in FIG. 9). Candidate group data having WCN data may be extracted.

In the present embodiment, when a one-to-one combination of WCN data and vehicle number is determined, it is determined as it is, but it can also be done as follows. When a one-to-one combination of WCN data and vehicle number is determined, a fixed flag is given. For the combination data to which this confirmation flag is added, if the same combination data as the combination data can be confirmed one-to-one from another candidate group data, the confirmation process is performed. This enables more reliable confirmation processing.

In the present embodiment, the candidate group data are processed in order, but it is also possible to preferentially process data with less combination data. This reduces the number of items to be compared with the combined data, so that the finalization process can be performed at an early stage.

C. Second embodiment (C1 about false detection)
In the above embodiment, a case has been described on the premise that there is a correct combination in the candidate group data, but it is also conceivable to extract an incorrect combination.

For example, when linking vehicle numbers based on the acquisition time of WCN data, even though there is no defect in each candidate group, correct information may not be included in any candidate group due to a time lag. This is because if the specifications of the DSRC antenna and camera are different, the detection timing will be different depending on the installation situation and the vehicle.

For example, assuming that FIG. 16A is correct candidate group data, if the timing of data acquisition and WCN acquisition by the camera is shifted, there is a risk of erroneous detection as shown in FIG. 16B. If such a deviation also occurs in other candidate group data, erroneous candidate combination data is extracted from a plurality of candidate group data, and as a result, it is determined.

(C2 erroneous detection prevention processing)
In order to avoid the above problem, the process shown in FIG. 18 is executed in this embodiment. A case where the candidate group data of FIG. 17A exists will be described below.

The CPU 23 reads each candidate group data and creates candidate combination data for each candidate group data (step S151 in FIG. 18). In this case, candidate combination data described above is created for all candidate group data K11 to K14. FIG. 19 shows the generated candidate combination data.

The CPU 23 deletes existing determined data for the created candidate combination data, if any (step S153 in FIG. 18). In this case, since existing fixed data does not exist, the CPU 23 sets the target group data number i=0 (step S155), and reads out all the candidate combination data of the 0th target group data (step S157). In this case, candidate combination data (N31, W30), (N31, W11), (N31, W32), (N11, W30), (N11, W11), (N11, W32), (N11, W30), (N11, W11), (N11, W32), (N32, W30), (N32, W11), (N32, W32) are read out.

The CPU 23 determines whether or not the candidate combination data of the attention group data is 1 combination data (step S159). In this case, since the 0th attention group data K11 is not 1 combination data, the CPU 23 sets the attention candidate number j=0 (step S163).

The CPU 23 reads the j-th candidate combination data (step S165). In this case, the 0th candidate combination data (N31, W30) is read. The CPU 23 performs score judgment for this candidate combination data (N31, W30) (step S167).

Score determination will be described with reference to FIG.

The CPU 23 determines whether or not there is candidate group data having the same WCN data "W30" as the candidate combination data (N31, W30) (step S301 in FIG. 21). In this case, since candidate group data K13 and K14 having WCN data "W30" exist, such candidate group data are extracted (step S302). The CPU 23 determines whether candidate group data including missing data is included in the extracted candidate group data (step S303). In this case, since the candidate group data K13 and K14 do not include a defect, the CPU 23 sets the first threshold (step S304a) and calculates the score of the candidate combination data (N30, W30) (step S305).

In the present embodiment, when missing data is not included, calculation is performed by (score=incidence rate=number of detected candidates for the same combination/number of detected WCN-containing candidate group data of interest).

For example, candidate group data in which candidate combination data (N31, W30) exists is zero. Therefore, the same combination candidate detection count is zero. On the other hand, there are a total of three candidate group data, K11, K13, and K14, which have combination data including the WCN data "W30" extracted in step S305. Therefore, the score is 0/3=0.

The CPU 23 determines whether or not the first threshold is exceeded (step S307). In this embodiment, the first threshold is set to 0.75. Therefore, the CPU 23 terminates the score determination process.

The CPU 23 determines whether or not the target candidate number j is the last (step S169 in FIG. 18). In this case, since the attention candidate number j is not the last, the CPU 23 increments the attention candidate number j (step S171), and reads the candidate combination data of j=1 (step S165). As a result, the next candidate combination data is read, and score determination processing (step S167) is performed in the same manner as described above.

In this case, the CPU 23 determines whether or not there is candidate group data having the same WCN data "W11" as the candidate combination data (N31, W11) (step S301 in FIG. 21). In this case, since there is no candidate group data having WCN data "W11", the score determination process ends.

The CPU 23 determines whether or not the target candidate number j is the last (step S169 in FIG. 18). In this case, since the attention candidate number j is not the last, the CPU 23 increments the attention candidate number j (step S171) and reads the candidate combination data of j=2 (step S165). As a result, the next candidate combination data is read, and score determination processing (step S167) is performed in the same manner as described above.

When the above processing is executed up to j=8, the judgment on the candidate combination data (N32, W32) is completed. As a result, the CPU 23 advances from step S169 to step S173 to determine whether or not all candidate group data have been examined.

In this case, since all the candidate group data have not been examined, the CPU 23 increments the target group data number i (step S175), and repeats the processing from step S157 onward.

Since the processing up to the score determination processing for the candidate combination data (N30, W30) of the candidate group data K13 is the same as the above, the description is omitted.

The CPU 23 extracts candidate group data having the WCN data "W30" of the candidate combination data (N30, W30) (step S301 in FIG. 21). In this case, since candidate combination data having the same WCN data "W30" exists in candidate group data K11 and K14, these candidate group data are extracted. The CPU 23 deletes candidate group data including missing data from the extracted candidate group data (step S303). In this case, nothing is deleted from the candidate group data K13 and K14 because they do not contain defects.

The CPU 23 calculates the score of the candidate combination data (N30, W30) (step S305).

In this case, there are two candidate group data, K13 and K14, in which candidate combination data (N30, W30) exist. Therefore, the same combination candidate detection count is two. On the other hand, there are three candidate group data, K11, K13, and K14, which have combination data including the WCN data "W30" extracted in step S301. Therefore the score is 2/3=0.66.

The CPU 23 determines whether or not the first threshold is exceeded (step S307). In this embodiment, since the first threshold is set to 0.75, the score determination process is terminated assuming that the first threshold is not exceeded.

Thereafter, the other candidate combination data of the candidate group data K13 and the candidate group data K14 are similarly judged, and in order not to exceed the first threshold value, if the detection result is as shown in FIG. can be prevented.

On the other hand, if the candidate group data shown in FIG. 17B exists, and the score of the candidate combination data (N30, W30) is calculated in the score determination process of step S305 in FIG. 21, 4/5=0.8. becomes. Therefore, the CPU 23 determines that the score of the candidate combination data (N30, W30) exceeds the first threshold value of 0.75, and determines whether or not it contradicts the existing fixed data (step S307 in FIG. 21). Then, if there is no contradiction, the combination is confirmed (step S311) and deleted from the object (step S313).

As described above, even if it can be uniquely identified by existing in a plurality of candidate group data, even if the candidate group data is detected by chance by considering the above score , is not determined unless there is a certain number of candidate group data.

In the above-described embodiment, the incidence rate is used as the score, but judging by such probability, if there is only one erroneous candidate group data, the incidence rate is 1/1=1, exceeding the threshold. There is a risk that the problem will occur. Therefore, in order to deal with such a case, if only the occurrence rate is used for determination, whether or not the number of candidate group data is at least two or more may be determined.

Moreover, even if the occurrence rate is the same, the probability of not being falsely detected increases depending on the number of candidate group data. That is, the occurrence rate is 1/2 in both the case where the same combination candidate detection number is 2 out of 4 and the case where the same combination candidate detection number is 5 out of 10. However, the latter is less likely to be wrong. Therefore, the threshold may be changed according to the number of detected same combination candidates.

Further, the number of same combination candidate detections may be weighted as described later. That is, (1) detection of the same candidate, which will be described later, the number of points is obtained by calculating 1 point*the number of candidate group data, and this number of points is used as the score.

Although the first threshold is set to 0.75 in this embodiment, it may be set arbitrarily depending on the camera, communication device, antenna, traffic volume, and the like.

Note that, in the case where the data conflicts with the existing fixed data in step S309 of FIG. 21, the data is regarded as a replacement candidate, as in the first embodiment.

(About C3 weighting)
As the score, the obtained incidence rate may be weighted in the determination of Condition 3 in the first embodiment, and the result may be used as the score. The reason for this is that similar events may occur under the same installation conditions and with the same vehicle at the same base, so it is thought that combining information from multiple bases will improve accuracy rather than information from the same base. be. Also, if the detection date and time zone are different, the traffic volume and the like will change even at the same base.

Specifically, for example, the calculated incidence rate may be multiplied by the following weight to obtain a score.

・(1) Detection of the same candidate, 1 point * number of candidate group data ・(2) Detection at different bases 0.5 points * number of bases ・(3) Detection at different dates and times 0.5 points * number of days In the case of Figure 17B, occurrence The rate was 0.8. So multiplying this by (4+0.5*3+0.5*4) gives a score of 0.8*(4+0.5*3+0.5*4)=6 points.

In this way, erroneous detection can be prevented by assigning points with different coefficients for (2) different bases and (3) different dates and times. In this case, a threshold may be appropriately set as the first threshold.

For example, the traffic volume at each base may be stored by date and time (for example, traffic volume is high during commuting hours from Monday to Friday, and low on Saturdays and Sundays), and the weighting may be changed according to the detected time. .

It should be noted that the traffic volume can also be determined by verifying the number of candidate group data per unit time, even if it is not stored separately.

(About C4 weighting and missing data processing)
As the score, the determination process of weighting the incidence rate and obtaining the score can be similarly applied to the defect-containing candidate group data if the threshold value is appropriately set.

For example, even though (N30, W30) is a correct combination, if the candidate group data shown in FIG. 20A exists, it may be detected as erroneous candidate group data such as candidate group data K20 and K21. be done. In this case, if the processing of condition 3 in FIG. 14 is performed as in the first embodiment, as shown in FIG. 20B, only one other candidate group data K22 is determined as incorrect combination data There is a risk that it will be done.

On the other hand, the score is determined by the above weighting, that is, (1) detection of the same candidate, 1 point * number of candidate group data, (2) detection at different sites 0.5 points * number of sites, (3) different Detection by date and time By adopting 0.5 points * number of days, it is possible to prevent erroneous decisions as follows.

For example, in the case of FIG. 20B, the incidence is 2/3=0.67. Therefore, multiplying this by the weight (2+0+0) gives a total of 0.67*(2+0+0)=1.33 points.

In contrast, in the case of FIG. 20C, the occurrence rate is 0.5. So multiplying this by (3+0.5*2+0.5*2) gives a total of 0.5*(3+0.5*2+0.5*2)=2.5 points.

Therefore, if the threshold value is set to 2 points, it is possible to confirm even the defect-containing candidate group data when it is detected at different bases and on different days, as in the latter case. Thus, even when there is defect-containing candidate group data, appropriate determination can be made by appropriately setting the threshold.

As a specific process in this case, if there is a defect in step S303 of FIG. 21, a second threshold lower than the first threshold may be set as the threshold (step S304b).

In addition, in the present embodiment, the WCN is focused on in the determination of step S305, but the vehicle number may be focused on.

When the first embodiment and the second embodiment are used together, the judgment of the first embodiment may be made after making the judgment of the second embodiment.

In this embodiment, the evaluation judgment value is obtained based on the occurrence rate itself and the weighted value thereof.

(D Third Embodiment)
Detection results at close times at the same location may be combined and processed in FIGS. 8 and 18 . For example, the candidate group data K11 and the candidate group data K12 in FIG. 17A are candidate group data that are located at the same site and have detection times that differ by one second. Therefore, candidate group data K11 and candidate group data K12 are combined to generate one candidate combination data. Candidate group data K11 is (N31, N11, N32)*(W30, W11, W32), and candidate group data K12 is (N30, N13)*(W31, W13). Therefore, the combined candidate group data is (N31, N11, N32, N30, N13)*(W30, W11, W32, W31, W13). Thereby, candidate combination data including the correct combination (N30, W30) can be generated. As a result, the correct combination can be determined not by the candidate combination data (N31, W30) but by the candidate combination data (N30, W30) common to the other candidate group data K13, K14.

In this manner, by storing the allowable time difference and merging candidate group data whose time difference is shorter than the allowable time difference at the same site, it is possible to prevent generation of correct candidate combination data. The permissible time difference may be stored.

In the above embodiment, the CPU 23 is used to realize the functions shown in FIG. 1, and this is realized by software. However, part or all of it may be implemented by hardware such as a logic circuit. Part of the processing of the program may be performed by an operating system (OS).

Claims (17)

A combination data determination system for determining a combination of WCN data and vehicle number obtained in association by vehicle information obtaining means,
A plurality of WCN data obtained from a vehicle by a vehicle information obtaining means are defined as WCN group data, and a plurality of vehicle number data detected in association with the WCN group data are defined as vehicle number group data, and the WCN group data and the vehicle number are defined as vehicle number group data. Candidate group data storage means for storing a plurality of candidate group data with different detection times or detection locations for candidate group data composed of group data;
Candidate combination data generation means for generating a plurality of combination data of WCN data of the WCN group data and vehicle number data of the vehicle number group data as candidate combination data for each of the candidate group data. ,
When it is assumed that certain candidate combination data among the candidate combination data is correct, if there is no contradiction in the relationship with other candidate combination data in the candidate group data including the assumed candidate combination data, the assumed candidate combination data combination determination means for determining candidate combination data;
A combined data decision system with In the combination data determination system of claim 1,
The combination determining means extracts other candidate group data having the same vehicle number or WCN data constituting the candidate combination data of interest from the plurality of candidate group data, and extracts the extracted candidate group. If all of the data have the candidate combination data of interest, confirming the candidate combination data;
A combined data determination system characterized by: In the combination data determination system of claim 2,
The combination determining means repeats the process of removing the determined candidate combination data and determining the assumed candidate combination data ;
A combined data determination system characterized by: In the combination data determination system of claim 1,
The combination determination means, when assuming that certain candidate combination data is correct,
When there are a plurality of combination data that do not cause contradiction in relation to other combination data in the candidate group data including the assumed candidate combination data , the candidate combination data are removed to determine the assumed candidate combination data. repeating the process of causing
A combined data determination system characterized by: In the combination data determination system according to any one of claims 1 to 3,
the combination determination means making the confirmation when there is no missing candidate combination data in the candidate group data;
A combined data determination system characterized by: In the combination data determination system of claim 4,
The combination determining means executes a process of removing candidate combination data when there are a plurality of combinations that do not cause contradiction when there is no missing candidate combination data in the candidate group data;
A combined data determination system characterized by: In the combination data determination system according to any one of claims 1 to 6 ,
If there is a contradiction with the candidate combination data that has already been determined, it is stored as data to be replaced.
A combined data determination system characterized by: A program for realizing, by a computer, combination data determination processing for determining a combination of WCN data and a vehicle number acquired in association by vehicle information acquisition means,
A combination data determination program for configuring the computer as the following means.
A plurality of WCN data obtained from a vehicle by a vehicle information obtaining means are defined as WCN group data, and a plurality of vehicle number data detected in association with the WCN group data are defined as vehicle number group data, and the WCN group data and the vehicle number are defined as vehicle number group data. Candidate group data storage means for storing a plurality of candidate group data with different detection times or detection locations for candidate group data composed of group data;
Candidate combination data generation means for generating a plurality of combination data of WCN data of the WCN group data and vehicle number data of the vehicle number group data as candidate combination data for each of the candidate group data. ,
When it is assumed that certain candidate combination data among the candidate combination data is correct, if there is no contradiction in the relationship with other candidate combination data in the candidate group data including the assumed candidate combination data, the assumed candidate combination data combination determining means for determining candidate combination data; In the combination data determination system according to any one of claims 1 to 3,
When the candidate group data includes missing candidate combination data, the combination determining means extracts candidate combination data without the missing from the candidate group data,
Other candidate group data having the same vehicle number or WCN data constituting the noted candidate combination data is extracted, and all of the extracted candidate group data have the noted candidate combination data. and obtaining a point based on the number of the extracted candidate group data, and determining the candidate combination data when this exceeds a threshold;
A combined data determination system characterized by: In the combination data determination system of claim 9 ,
adjusting the points so that the points are higher when the extracted candidate group data differ in detection date or detection location;
A combined data determination system characterized by: In the combination data determination system of claim 9 or claim 10 ,
when the number of missing candidate combination data included in the candidate group data is less than a predetermined number, adjusting the points so as to be higher ;
A combined data determination system characterized by: In the combination data determination system of claim 1,
The combination determining means is
With respect to the plurality of candidate group data, other candidate group data having the same vehicle number or WCN data constituting the focused candidate combination data are set as first candidate group data, and the number thereof is obtained, and Calculating the number of candidate group data containing the same candidate combination data as the noted candidate combination data among the first candidate group data as second candidate group data ,
Obtaining an evaluation judgment value based on a value obtained by dividing the number of the second candidate group data by the number of the first candidate group data,
determining candidate combination data when the evaluation judgment value exceeds a pre-designed threshold;
characterized by The combined data determination system of claim 12 ,
When obtaining the evaluation judgment value based on the number of the second candidate group data , the combination determining means increases the evaluation judgment value if there is second candidate group data different in detection date or detection location. and
determining the candidate combination data when the evaluation judgment value exceeds the threshold;
characterized by In the combination data determination system of claim 1,
The combination determining means unites candidate group data at close times at the same base to form one candidate group data;
characterized by In the program of claim 8,
The combination determining means is
Among the plurality of candidate group data, the number of other candidate group data having the same vehicle number or WCN data constituting the focused candidate combination data is obtained as a first number of candidate group data, and determining the number of candidate group data containing the same candidate combination data as the noted candidate combination data among the candidate group data as a second number of candidate group data;
Obtaining an evaluation judgment value based on a value obtained by dividing the second candidate group data number by the first candidate group data number;
determining candidate combination data when the evaluation judgment value exceeds a pre-designed threshold;
characterized by In the program of claim 15,
When obtaining the evaluation judgment value based on the number of the second candidate group data, the combination determination means increases the evaluation judgment value if there is the second candidate group data having a different detection date or detection location. and
determining the candidate combination data if the adjusted evaluation judgment value exceeds a threshold;
characterized by In the program of claim 8,
The combination determining means unites candidate group data at close times at the same base to form one candidate group data;
characterized by

Images