TWI579809B - Vehicle ID Code Matching Method - Google Patents

Description

Vehicle identification code matching method

The invention relates to an identification code matching method, which uses function values to represent the correspondence degree between two identification codes stored on the same item, and thereby repeatedly trains and establishes two identification code correspondence tables between multiple items, by One of the identification codes identifies another identification code that most likely matches the same item.

At present, the domestic highway electronic toll collection system uses the original infrared DSRC to adopt RFID technology as the calculation basis for vehicle tolls. The RFID tag (the remote power collection is named eTag) has more than 6 million users due to its size advantage. There will be great business opportunities.

In the case where the vehicle information is not disclosed, the RFID reader and the license plate identification can obtain the RFID tag identification code and the license plate number, respectively, but the relationship between the two cannot be confirmed.

General license plate recognition system, due to lighting, license plate color, fouling and other factors, the possibility of identification errors is inevitable, so the use of the license plate number as the vehicle's unique identification code, in addition to the legality of the fake license plate, There are also many technical problems to improve the recognition rate.

The matching method between common license plate recognition and other identification codes is mainly carried out by the time difference between the data input time of the two, which is to observe the reading or triggering time of the two kinds of data synchronously. If the time difference between the two is smaller, it can be regarded as the same. Vehicles, however, in the case of multi-lane free flow, it is bound to cause a match Difficulties, and if necessary, will increase the cost of subsequent manual identification.

In order to solve the above problems, the integration of these two technologies is a key technology that is innovative and helps to improve the performance of the vehicle identification system.

The invention provides an identification code matching method for an item having two identification codes, wherein a function value is used to represent the correspondence degree between two identification codes, and thereby establishing two identification code correspondence tables between a plurality of articles, From one of the identification codes, another identification code that most likely matches the same item can be found.

The invention is applied to identify the RFID tag identification code and the license plate number of the vehicle. In the case of unknown vehicle data, the administrator reads the RFID tag identification code of the vehicle through the RFID system, and can find the best corresponding by means of table lookup. The license plate number is confirmed and the consistency with the license plate number generated by the license plate recognition, and the result of the license plate recognition can also be used as the training material of the identification code correspondence table.

It is assumed that if an item has two types of identification codes A and B and there is a system having the function of reading the A and B identification codes of the items in a sensing range, the following will take A as the main identification code and B as the corresponding identification code. Explain, and vice versa.

When there are multiple items simultaneously passing through the sensing reading range of the system of the present invention, for example, the sensing column sensing range of the eTag on the national road, the system will detect two identification codes of the same item according to different detection methods, such as on the national road. captured by the imaging equipment, automobiles and the license plate number to the RFID reader eTag identification result, the system will generate different m a pen type identifier (license plate number) and B n pen identification code (recognition result eTag), wherein The class A identification code generated by the system at time point t _{A,i is} defined as ID _A , and the class A identification code of the i-th pen is ID _A,i , and the time of the system is t _A,i ± T _D,MAX The number of pens of the B-type identification code generated inside is defined as n _i , and the J-th class B identification code read by the system at the time point t _B,j is defined as ID _B,j , j =1... n _i And n _i n , the system generates ID _{A, i} and ID _{B, j} time difference must meet the conditions The above abs ( ) is an absolute value function, and T _{D, MAX is} the maximum time difference between the type A identification code and the class B identification code that the system of the present invention allows to read the same item.

The invention can establish an identification code correspondence table by using a batch or a continuous update manner, and the identification code correspondence table is used for finding a matching relationship between the class A identification code of the object and the class B identification code, wherein a column A class The matching relationship between the identification code and the class B identification code includes data fields such as ID _A , ID _B , T _B , C _B , W, etc., wherein ID _A is a class A identification code, and ID _B is a class B identification code; T _{C, C B,} W and other data fields match the parameters for this ID _a and ID _B of wherein, T _B of this input time ID _B, C _B is the cumulative number of inputs ID _B (initial value 0 for this ), W is the matching proportion of ID _A and ID _B ; when the system reads input ID _A and ID _B , the data field setting rules of the matching relationship on the identification code correspondence table are respectively as follows ( 2)~(6): ID _A = ID _{A, i} .... formula (2)

ID _B = ID _B,j ....Formula (3)

T _B ( k )= t _B,j ....Formula (4)

C _B ( k )= C _B ( k -1)+1....formula (5)

W ( k )= f _w ( t _A,i ,T _B ( k ) , C _B ( k )).... (6) where k is the same ID _{A, i} and ID _{B, j} match relationship Enter (including new or updated) the sequence number, then f _w is a function of t _A,i , T _B ( k ), C _B ( k ).

Compared with the prior art, the invention has the following better effects:

RFID reader and license plate when the vehicle information is not disclosed The identification can obtain the RFID tag identification code and the license plate number separately, but the relationship between the two cannot be confirmed. The present invention can establish the correspondence between the two identification codes by batch or instant data input.

The general license plate recognition system has the possibility of identifying errors due to many factors such as image resolution, illumination, license plate color, and fouling. Therefore, the license plate number is used as the unique identification code of the vehicle, except for the legality of the fake license plate. There are technical problems to improve the recognition rate. If the present invention is used to establish a matching relationship between the RFID tag and the license plate, the best corresponding license plate number of the vehicle with the RFID tag can be provided, thereby reducing the false alarm caused by the license plate identification error.

The matching between common license plate recognition and other identification codes is mainly based on the time difference between the data input time. The reading or triggering time of the two data is synchronized. If the time difference between the two is smaller, it is regarded as the same part. Vehicles, however, in the case of multi-lane free flow, will cause difficulties in matching and will increase the cost of manual identification if necessary. The invention can obtain the best matching license plate number through the training of continuous historical data to accelerate the processing of the non-login customer vehicle.

S101~S116‧‧‧Step procedure

S201~S208‧‧‧Step procedure

3011‧‧‧First front-end device group

3011A‧‧‧RFID reader

3011B‧‧‧ License Plate Image Identification System

3012‧‧‧Second front-end device group

3012A‧‧‧RFID reader

3012B‧‧‧ License Plate Image Identification System

302‧‧‧First Communication Interface

303‧‧‧Control server group

304‧‧‧Second communication interface

305‧‧‧ID code correspondence table database

3051‧‧‧eTag EPC and car number correspondence table

3052‧‧‧eTag EPC and mobile phone number correspondence table

306‧‧‧Application Server

4011‧‧‧Base Station

4012‧‧‧Base Station

402‧‧‧Mobile Communication Network Application Server

FIG. 1 is a schematic diagram showing a flow of establishing a correspondence table of A and B identification codes in a method for matching a vehicle identification code according to the present invention.

2 is a schematic diagram of a query flow of a correspondence table of A and B identification codes in a method for matching a vehicle identification code according to the present invention.

FIG. 3 is a schematic structural diagram of a system architecture of a first embodiment of a method for matching a vehicle identification code according to the present invention.

4 is a schematic diagram of a system architecture of a second embodiment of a vehicle identification code matching method according to the present invention.

The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. .

The process of establishing the correspondence table of the A and B identification codes in the vehicle identification code matching method of the present invention is as shown in FIG. 1. The process of step S101 is started. After the process of establishing the identification code correspondence table is started, the process proceeds to step S102 to wait for the identification code to be input. in batch mode or real time waiting for the identification code input, i.e. input a except the pen-based identification code m and B n pen identification code, in order to establish the matching relationship of such class a and class B PIN identification code acquired in step S103 The i- type A-type identification code and the corresponding-numbered B-type identification code, the i-th type A-type identification code is ID _{A, i} , and the corresponding-number B-type identification code is n _i- pen B-type identification code, Where the n _i pen B class identification code must satisfy the formula (1) abs ( t _{A, i} - t _{B, j} ) T _D,MAX , and thus the n- _th pen B-type identification code proceeds to step S104 to obtain the j-th B-type identification code, the j-th B-type identification code is represented as ID _B,j , and then the inspection step S105 is performed. Whether there is a matching relationship between the i-th class A identification code ( ID _{A, i} ) and the j-th class B identification code ( ID _{B, j} ) in the correspondence table, and if yes, proceeding to step S106 to update the existing existence of the identification code correspondence table The matching relationship is mainly to update the two data fields of T _B and C _B ; if the result of the determination in step S106 is no, then step S107 is added to add a matching relationship to the identification code correspondence table, wherein the main establishment does not include the matching proportion W column. Other parameters of the bit.

As shown in FIG. 1, it is confirmed whether the matching relationship of all the class B identification codes ( ID _B ) is completed in step S108 (the field information is not yet complete), and if not, the process returns to step S104 to obtain the next If the result of step S108 is YES, then the calculation and update of the matching proportion is performed, and the matching relationship whose matching proportion is lower than a certain degree is deleted, so as to avoid unrestricted growth of the data volume of the corresponding table; the program is the first step S109 queries all the matching relationship sets in the correspondence table that include the i-th A-type identification code ( ID _{A, i} ), the matching relationship set is represented as S _i , and then the step S110 takes out the z- _th in the matching relationship set ( S _i ) a pen matching relationship, the z- th pen matching relationship is expressed as M _z , and the matching weight W of the M _z in the corresponding table is updated, and then the confirmation step S111 performs the z- th matching relationship to satisfy the deletion condition, and the deletion condition is set to L ₁ if yes, the z pen step S112 to delete the matching relationship of the correspondence table is performed; if not, step S113 updates the first pen z matching relationship of the correspondence table; then, performs step S114 to confirm whether the matching is completed off Set (S _i) of all updates matching relationship in or deleted, if not, returns to step S110 to obtain the next piece of data, if the step S114 is yes, then step S115 is completed to establish all Class A identification code (ID _A) matches If the result of step S114 is NO, the process returns to step S103 to obtain the next data, that is, the class B identification code of the n _i pen satisfying the formula (1); if the result of step S115 is YES, the process proceeds to step S116.

The query flow of the identification code correspondence table is as shown in FIG. 2 . First, in step S201, the identification code correspondence table query process starts, and the A-type identification code to be checked is obtained. In step S202, the identification code i-th A-type identification code ( ID _{A is} input. _{, i} ), the identification code ID _A is the basis of the query identification code correspondence table, and then the ID _A field of the identification code correspondence table is queried to find all the records including the ID _{A, i to} be checked, that is, proceeding Step S203: querying the identifier correspondence table and obtaining a matching relationship set ( S _i ) including the ith type A identification code, and then executing S204 to obtain the class B identification code having the largest proportion in the matching relationship set, in order to determine the used Whether the matching relationship is correct, whether the execution determining step S205 meets an access condition, the access condition is set to L ₂ , and if so, the matching relationship has a high correct rate, and the recommendation uses the matching relationship with the correct rate of step S206; If no, the matching relationship has a low accuracy rate, and it is not recommended to use the matching relationship with the correct rate in step S207. Finally, the process of completing the query process is completed.

An embodiment of the present invention is applied to an in-vehicle device identification code and a license plate number. In the prior art, a system for recognizing a license plate using a photographic surveillance image may cause an error rate due to various factors such as image resolution, illumination, license plate color, and stain. Highly, general vehicle identification applications, such as vehicle trading systems implemented in national electronic toll collection systems, use short-range communication systems with high success rates, such as Dedicated Short-range Communications (DSRC) or RF. The Radio Frequency Identification (RFID) system is supplemented with an image license plate recognition system when necessary to check the vehicles of the Infrared On Board Unit (eTag) without the short-range communication system to ensure all vehicles (including the vehicle) If the equipment fails or the vehicle is not logged in, it can be correctly identified and traded by the system. At this time, the system needs to match the vehicle equipment identification code of the communication system with the vehicle number of the license plate recognition system to separately filter the logged in and not logged in. Vehicles for subsequent different payment operations. The general matching method is to determine whether the two are attributed to the same vehicle by the time difference between the in-vehicle device identification code and the license plate identification data input time at the same place. If the time difference between the two is less than a certain value, the same vehicle is determined, but Even if the short-distance communication system and the license plate image recognition system are time-synchronized, the time for reading or triggering between subsystems is different. On the practical side, in the case of multi-lane free flow, adjacent vehicles pass through or cross the lane at the same time. Induction can make system matching difficult, and increase the workload of manually identifying unmatched parts.

The focus of the present invention is to give an in-vehicle device identification code a matching proportion with a certain license plate number by a function including the number of readings and the reading time. Therefore, when a certain vehicle passes through the system, the system sets its in-vehicle device identification code. The matching relationship is established with the license plate numbers of multiple vehicles passing at the same time. After multiple sensing and matching, the proportion of the correct matching relationship will be increased. Therefore, all the matching relationships of the single in-vehicle device identification codes are sorted by matching proportions. After that, the best matching car number with the highest specific gravity will be obtained. If the present invention is applied to a national highway electronic toll collection system, the in-vehicle device adopts an EPC (Electronic Product Code) outer code using an RFID system as an example of the present invention. The class A identification code and the license plate number in the matching method are regarded as the class B identification code, and the system architecture thereof is as shown in FIG. 3, and includes a plurality of front end device groups (by the first front end device group 3011 and the second front end device group). Group 3012 indicates that more front-end device groups can be extended, which are multiple RFID readers installed in different locations (in the figure, RFID readers 3011A, R) The FID reader 3012A is represented by a plurality of license plate image recognition systems (indicated by the license plate image recognition system 3011B and the license plate image recognition system 3012B), and the front end device group is transmitted through a first communication interface 302 (for example, an Ethernet interface). Connected to the control server group 303, the control server group may include a plurality of control servers, and the control server sets and controls multiple RFID readers and multiple license plate images in the front-end device group The identification system, the control server group is connected to the identification code correspondence table database 305 through a second communication interface 304, and the matching method of the eTag identification code and the license plate number is implemented on the control server group 303, and the matching method is The identification code establishing program will establish an eTag EPC and car number correspondence table 3051 on the identification code correspondence table database 305; in addition, an application server 306 (for example, an electronic charging application server) is used to set the front end device and obtain the transaction identification code. Corresponding to the vehicle image for the corresponding payment process, through the communication interface 304 and the control servo The device group 303 and the identification code correspondence table database 305 are connected. After the vehicle passes, the application server 306 can query the identification code correspondence table database 305 to obtain the matching relationship between the eTag EPC and the vehicle number.

In this embodiment, the detailed parameters of the matching method are set as follows: setting formula (1) abs ( t _{A, i} - t _{B, j} ) T _{D, MAX} T _{D, MAX} is 3 seconds, which means that the eTag input and the license plate identification result input have a maximum time difference of 3 seconds, which can be adjusted by the system operator according to the actual operating environment requirements of the system; formula (2)~ (5) as follows: ID _A = ID _{A, i} .... formula (2)

ID _B = ID _B,j ....Formula (3)

T _B ( k )= t _B,j ....Formula (4)

C _B ( k )= C _B ( k -1)+1....Formula (5) is the algorithm for setting the data movement; formula (6): W ( k )= f _w ( t _A,i ,T _An example of a matching specific gravity function f _W in _B ( k ) , C _B ( k )) is equation (7): , where f _C is the cumulative input number function of the class B identification code ID _B in the matching relationship, and its definition is designed as equation (8): f _C =1+ log ( C _B ), where C _{B is the} matching relationship The cumulative input number of the class B identification code ID _B,j (the input sequence number k is omitted below), log ( C _B ) is the logarithm function of the logarithm of C _B on the basis of 10, and f _T is the class A of the matching relationship The input time difference between the input code of the identification code ID _A and the class B identification code ID _B is defined as the formula (9): f _T =1+ abs ( t _A,i - T _B ), where t _A,i is The input time of the i- th class A identification code ID _{A, i} , T _B is the input time of the class B identification code ID _B in the matching relationship, and abs ( t _{A, i} - T _B ) is the above two time The absolute value of the time difference, and t _A,i and T _B are floating point values in days.

In the formula (1), the larger T _{D, MAX} allows the data input time difference generated by the slower speed, but it may also generate more matching relationship when the traffic volume is large, and increase the data volume of the identification code correspondence table. In contrast, the identification code correspondence table establishment process has a deletion condition L ₁ , which is an active restriction condition for controlling the data volume of the identification code correspondence table, and a looser condition will make the correspondence table retain more matching relationships, and is more strict. The condition that causes the matching relationship of vehicles with rare usage is deleted by the system due to the small number of input codes or long intervals. For example, the constraint condition L ₁ can be designed as Equation (10): SizeOf ( S _i )>5, A maximum of 5 matching relationships with the highest proportion of matching are retained, where SizeOf ( S _i ) is the number of matching relationships for all matching relationship sets S _i including this ID _{A, i} , and where S _i has been sorted according to the matching proportion.

On the other hand, if you want to query, you can filter the cumulative number of times more than 10 times and the previous input time is within 30 days. ), the access condition L ₂ in the query flow is set to equation (11): C _B 10 and W W _{ACCU, LDATE} When updating the identification code correspondence table in an immediate and continuous manner, it is necessary to pay attention to whether the update matching proportion is included when querying the matching relationship, because the system does not read the ID _A and ID _B time by the front end identification code input device. do the same, so the system is not sure whether the other identification code is present in the ID _B t _a, ID _a case in order to match the query matches the ID _B gravity W, the proportion may be prior match W (k -1) ID _B of It may also be the matching weight W ( k ) of this ID _B. However, if the vehicle frequently and repeatedly appears, the single match cannot directly dominate the matching weight. In this case, the matching weight W ( k ) or W ( k -1) will not affect the matching result. When one of the specific gravity is used, the following scheme can be adopted: (1) The system delay can be used until a time t _{D 1} and t _{D 1} > t _B + T _{D, MAX} , and then the matching parameters are updated to ensure that t _A The matching weight of the query is still the matching weight W ( k -1) of the previous ID _B , (2) if the system can ensure t _A > t _B or delay to a time t _{D 2} and t _{D 2} > t _A + T _{D, MAX} then query, then the matching weight of the query is the matching weight W ( k ) of the ID _B.

The second embodiment of the present invention is applied to establish an eTag user database under the premise of no eTag login user data, and the system provider can establish two identification code correspondence tables, and one is an eTag EPC to car number correspondence table (such as the first embodiment). The second is the eTag EPC to the mobile phone number (or IMEI) correspondence table, the system architecture of the application system using the present invention is shown in FIG. 4, and the application system includes the system architecture except the first embodiment, and the control server group. The group 303 is further connected to the mobile communication system application server 402 through the communication interface 304, and thereby obtains the user mobile phone number data of the plurality of base stations (represented by the base station 4011 and the base station 4012 in the figure) covering the plurality of locations, and is controlled. On the server group 303, a matching method between the eTag identification code and the mobile phone door number is also implemented. The identification code establishing program of the matching method establishes an eTag EPC and mobile phone door number correspondence table 3052 on the identification code correspondence table database 305. For example, the location of the installation of the system RFID reader 3011A belongs to a mobile communication base station 4011, and the EPC of the eTag is used as the class A identification code, and the door number of the mobile phone can be used. As the Class B identification code; when the user's vehicle passes the RFID reader 3011A, the system queries all the mobile phone door numbers in the coverage area of the mobile communication base station 4011 at that time, and the correspondence table between the eTag identification code and the mobile phone door number. The matching relationship between the EPC of this car eTag and all the door numbers is established.

Referring to FIG. 4 , the maximum time difference T _D,MAX , the matching weight function f _W and the deletion condition L ₁ can be set according to the first embodiment. After the eTag identification code is trained on the mobile phone door number correspondence table, the EPC is matched correctly. The matching weight of the mobile phone number will gradually increase until it is greater than the matching weight matching to other incorrect mobile phone numbers, so the system will be able to retrieve the correct mobile phone number. By combining the two identification code correspondence tables, a complete eTag user database can be established, and the application server 306 can include the eTag identification code, the license plate number, the user mobile phone door number according to the application query, and then apply the action gate number. Customer contact method and account information.

The detailed description of the preferred embodiments of the present invention is intended to be construed as the invention The patent scope of this case.

In summary, the present invention is innovative in terms of technical ideas, and also has various functions that are not in the prior art, and has fully complied with the statutory invention patent requirements of novelty and progressiveness, and has filed a patent application according to law, and invites you to approve the invention. The patent application was inspired to invent, and it was a matter of feeling.

S101~S116‧‧‧Step procedure

Claims (5)

A vehicle identification code matching system, comprising: at least one identification code detecting device group, each of the identification code detecting device groups respectively respectively comprising a plurality of identification code input devices, each of the identification code input devices generating at least one identification code input data a first communication interface, the first communication interface is coupled to the identification code detecting device group; at least one control server is connected to the first communication interface and controlled by the first communication interface The identification code input device is configured to generate an identification code correspondence table establishing program; a second communication interface, the second communication interface is coupled to each of the control servers for transmitting a message sent by each control server to a database; The database is connected to the second communication interface, and the database generates at least one identifier correspondence table according to the identifier correspondence table establishment program and the identifier input data, and the database continuously updates the identifier correspondence table. Calculate the weight between the identification code input data. The vehicle identification code matching system of claim 1, wherein the second communication interface is coupled to a mobile communication application server, and the mobile communication application server inputs the identifications in the identification code detection device groups. The code input data is used to query mobile phone information of at least one mobile communication base station tube in the communication range of the identification code detecting device group. The vehicle identification code matching system of claim 1, wherein the second communication interface is coupled to an application server, and the application server inputs the identification code input data when the identification code detecting device group inputs The query covers the identification code correspondence table in the database to find other kinds of identification codes corresponding to the identification code. A vehicle identification code matching method includes the following steps: querying an identification code correspondence table according to a class A identification code registered by an item, so as to find a correct one of the items whose matching specific value is the maximum value in the identification code correspondence table. a class B identification code; wherein the identification code correspondence table stores an input time including the class A identification code, the complex class B identification code, each class B identification code, and the cumulative input number of each class B identification code, the A a matching proportion of the class identification code and each of the class B identification codes, wherein the matching specific gravity is set according to a matching specific gravity function, wherein an input parameter of the matching specific gravity function includes an input time of the class A identification code, and each of the class B identification codes The input time, and the cumulative number of inputs of each of the Class B identification codes. The vehicle identification code matching method according to claim 4, wherein the matching specific gravity function is Where f _C is a cumulative input number function of each of the Class B identification codes in the matching relationship, and f _T is an input time difference function of the input time of the Class A identification code in the matching relationship and each of the Class B identification codes.