CN112115946A

CN112115946A - License plate fake-license plate identification method and device, storage medium and electronic equipment

Info

Publication number: CN112115946A
Application number: CN202011023670.7A
Authority: CN
Inventors: 不公告发明人
Original assignee: Chongqing Unisinsight Technology Co Ltd
Current assignee: Chongqing Unisinsight Technology Co Ltd
Priority date: 2020-09-25
Filing date: 2020-09-25
Publication date: 2020-12-22
Anticipated expiration: 2040-09-25
Also published as: CN112115946B

Abstract

The embodiment of the application provides a method and a device for identifying fake license plates of license plates, a storage medium and electronic equipment, and relates to the field of license plate identification. The method comprises the following steps: acquiring a snapshot sequence of a target license plate; the snapshot sequence comprises a plurality of bayonet identifications, one bayonet identification corresponds to one snapshot time, and the plurality of bayonet identifications are arranged according to the sequence of the snapshot times; acquiring all target gate identification groups in a snapshot sequence; and determining the number of the fake plate of the target license plate according to the number of the target checkpoint identification groups. In a period of time, if the number of the target bayonet identification groups exceeds 1, it is indicated that a plurality of vehicles with the license plates being the target license plates run in the period of time, and therefore the number of the fake plate of the target license plate can be determined according to the number of the target bayonet identification groups, so that the number of the fake plate can be effectively identified, and the accuracy of fake plate identification of the license plate is improved.

Description

License plate fake-license plate identification method and device, storage medium and electronic equipment

Technical Field

The application relates to the field of license plate recognition, in particular to a method and a device for recognizing fake license plates of license plates, a storage medium and electronic equipment.

Background

The fake-license vehicle is characterized in that fake-license plates with the same number plate are used on other vehicles according to real license plates.

The existing fake plate identification method utilizes a large number of gate cameras in a road network to capture passing vehicles, identifies the license plates through a license plate identification algorithm, carries out rule limitation and related mining on the license plate passing space-time track, finds the irrational property of the license plate running track, and accordingly judges whether the license plates are fake plate or not and becomes an effective means for detecting fake plate vehicles.

However, the conventional fake plate identification method cannot identify the number of fake plates when identifying fake plates, and has poor identification effect.

Disclosure of Invention

The application aims to provide a license plate fake-license plate identification method, a license plate fake-license plate identification device, a storage medium and electronic equipment, and the license plate fake-license plate identification method, the license plate fake-license plate identification device, the storage medium and the electronic equipment can effectively identify the number of fake-license plates and improve the license plate fake-license plate identification accuracy.

The embodiment of the application can be realized as follows:

in a first aspect, an embodiment of the present application provides a method for identifying a license plate of a license plate, including:

acquiring a snapshot sequence of a target license plate; the snapshot sequence comprises a plurality of bayonet identifications, one bayonet identification corresponds to one snapshot time, and the plurality of bayonet identifications are arranged according to the sequence of the snapshot times;

acquiring all target gate identification groups in the snapshot sequence;

each target bayonet identification group comprises a plurality of target bayonet identifications which are arranged according to the sequence of the snapshot time, and any two adjacent target bayonet identifications meet the preset condition; the preset condition is characterized in that the difference value of two snapshot times corresponding to any two adjacent target bayonet identifications is greater than or equal to the shortest passing time between two bayonets corresponding to any two adjacent target bayonet identifications;

and determining the number of the fake plate of the target license plate according to the number of the target checkpoint identification groups.

In a second aspect, an embodiment of the present application provides a fake plate identification device for a license plate, including:

the acquisition module is used for acquiring a snapshot sequence of the target license plate; the snapshot sequence comprises a plurality of bayonet identifications, one bayonet identification corresponds to one snapshot time, and the plurality of bayonet identifications are arranged according to the sequence of the snapshot times;

the recognition module is used for acquiring all target bayonet identification groups in the snapshot sequence;

the identification module is further used for determining the number of the fake plate of the target license plate according to the number of the target bayonet identification groups.

In a third aspect, the present application provides a storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the steps of the method of any one of the foregoing embodiments.

In a fourth aspect, an embodiment of the present application provides an electronic device, which includes a processor and a memory, where the memory stores a computer program, and the processor implements the steps of the method in any one of the foregoing embodiments when executing the computer program.

The beneficial effects of the embodiment of the application include: each target bayonet identification group comprises a plurality of target bayonet identifications which are arranged according to the sequence of the snapshot time, and any two adjacent target bayonet identifications meet the preset condition; and the preset condition is characterized in that the difference value of two snapshot times corresponding to any two adjacent target bayonet identifications is greater than or equal to the shortest passing time between two bayonets corresponding to any two adjacent target bayonet identifications. Therefore, a plurality of target checkpoint markers in the target checkpoint marker group actually represent a space-time trajectory sequence of the target license plate. And in a period of time, if the number of the target bayonet identification groups exceeds 1, it is indicated that a plurality of vehicles with target license plates all run in the period of time, therefore, the number of the fake license plates of the target license plates can be determined according to the number of the target bayonet identification groups, so that the number of the fake license plates can be effectively identified, and the accuracy of fake license plate identification of the license plates is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.

Fig. 1 is a block diagram of an electronic device according to an embodiment of the present disclosure;

fig. 2 is a schematic flow chart of a license plate fake-license plate identification method according to an embodiment of the present disclosure;

FIG. 3 is a flowchart illustrating a possible implementation of S200 of the license plate fake-license recognition method shown in FIG. 2;

FIG. 4 is a flowchart illustrating a possible implementation of the license plate fake-license identifying method S210 shown in FIG. 2;

FIG. 5 is a flowchart illustrating a possible implementation of the license plate fake plate identification method S211 shown in FIG. 4;

FIG. 6 is a flowchart illustrating a possible implementation of S210 of the license plate fake-license recognition method shown in FIG. 2;

fig. 7 is a schematic flowchart illustrating a process of performing the first hanging bayonet determining step S210B according to an embodiment of the present application;

fig. 8 is a schematic flowchart illustrating the first hanging bayonet determining step S210C according to an embodiment of the present disclosure;

fig. 9 is a flowchart illustrating the step S210D of executing the abnormal bayonet determination according to the embodiment of the present application;

fig. 10 is a schematic flowchart illustrating the second suspension bayonet determining step S210E according to an embodiment of the present application;

fig. 11 is a schematic view illustrating a situation that a road network of a license plate passes through a gate in the embodiment of the present application;

fig. 12 is another schematic view illustrating a situation that a road network of a license plate passes through a gate in an embodiment of the present application;

fig. 13 is a functional block diagram of a license plate fake-license plate recognition apparatus according to an embodiment of the present disclosure.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations.

Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.

Furthermore, the appearances of the terms "first," "second," and the like, if any, are used solely to distinguish one from another and are not to be construed as indicating or implying relative importance.

It should be noted that the features of the embodiments of the present application may be combined with each other without conflict.

In the process of implementing the technical solution of the embodiment of the present application, the inventors of the present application find that:

the fake-license vehicle is characterized in that fake-license plates with the same number plate are used on other vehicles according to real license plates. At present, illegal personnel can generally select the vehicle license plate of the same brand, the same motorcycle type, the same colour to carry out the fake plate when carrying out fake plate, and under this kind of condition, it is difficult directly to discover whether normal through outward appearance and license plate contrast mode. Therefore, when the license plate is identified in a fake plate mode, the traditional information comparison based on vehicle registration is not applicable.

With the maturity of license plate recognition technology, the license plate recognition precision reaches a higher level, a large number of gate cameras in a road network are used for capturing passing vehicles, license plates are recognized through a license plate recognition algorithm, the space-time trajectory of license plate passing is subjected to rule limitation and relevant mining, the irrational nature of the license plate running trajectory is found, and therefore the possibility that whether the license plates are subjected to fake plate is judged to be an effective means for detecting fake plate vehicles.

In the scheme of detecting the fake-licensed cars by using the car passing snapshot data of the bayonet, the judgment rule and the judgment method generally have two modes: firstly, based on bayonet connectivity judgment, forming a spatial track sequence for a snapped license plate according to snapping time, analyzing a bayonet sequence of the license plate continuously passing through by combining a pre-set road network bayonet direct communication relation graph, and judging whether the connectivity of each bayonet point of the license plate is reasonable, thereby judging whether the license plate is a fake license plate; and secondly, based on space-time rationality judgment, forming a space track sequence for the snapshot license plates according to snapshot time, judging whether the vehicles corresponding to the same license plates pass between an upstream bayonet and a downstream bayonet reasonably to judge whether the license plates are overlapped or not by analyzing continuous passing records of the same license plates, wherein the judgment can be based on whether the passing time is less than the minimum time required by the two bayonets or not, and whether the average speed is greater than the maximum speed of the two bayonets or not.

The existing license plate fake-licensed identification method only judges the fake-licensed vehicle based on the bayonet connectivity and the bayonet traffic space-time rationality, and has three problems which cannot be avoided:

firstly, the fake-licensed vehicle identification method is extremely sensitive to license plate identification errors, because of the influence of factors such as unreasonable camera layout parameters (height, horizontal angle, inclination angle and the like), stained license plates, poor shooting physical conditions, special unsuitability for shooting weather due to rain and snow and the like, license plate identification has certain probability of errors (the empirical value is about 2%), the error identification license plates are probably the same as the normal running license plates of a road network if being larger, but the spatial position is far away, the phenomenon makes the unreasonable situation extremely obvious and becomes the most important factor influencing the identification precision of the fake-licensed vehicle;

secondly, the camera has the possibility of multi-shot missing shooting, and the judgment of the bayonet connectivity and the space-time rationality can be influenced, especially the missing shooting condition.

Thirdly, the passing space-time rationality of the bayonets based on the average speed is judged, the distance between the two bayonets needs to be calculated, the more accurate distance calculation needs to be conducted according to the distance of a navigation path instead of the linear distance between longitude and latitude coordinates, professional map navigation is often needed, and the front-end camera snapshot result is often in a private network and does not have the internet, so that the front-end camera snapshot result is difficult to obtain.

Therefore, the conventional fake plate identification method cannot identify the number of fake plates when identifying fake plates, and has poor identification effect.

To address the deficiencies noted in the background and discovery by the inventors, embodiments of the present application provide an electronic device 110. Fig. 1 is a block diagram of an electronic device according to an embodiment of the present disclosure. The electronic device 110 may include a memory 111, a processor 112, a bus, and a communication interface, the memory 111, the processor 112, and the communication interface being electrically connected to each other, directly or indirectly, to enable transmission or interaction of data. For example, the components may be electrically connected to each other via one or more buses or signal lines.

The processor 112 may process information and/or data related to the method for license plate fake-license identification to perform one or more functions of the method for license plate fake-license identification described herein. For example, the processor 112 may: acquiring a snapshot sequence of a target license plate; the snapshot sequence comprises a plurality of bayonet identifications, one bayonet identification corresponds to one snapshot time, and the plurality of bayonet identifications are arranged according to the sequence of the snapshot times; acquiring all target gate identification groups in a snapshot sequence; each target bayonet identification group comprises a plurality of target bayonet identifications which are arranged according to the sequence of the snapshot time, and any two adjacent target bayonet identifications meet the preset condition; the preset condition is characterized in that the difference value of two snapshot times corresponding to any two adjacent target gate identifiers is greater than or equal to the shortest elapsed time between two gates corresponding to any two adjacent target gate identifiers; and determining the number of fake-licensed target license plates according to the number of the target checkpoint identification groups, thereby realizing the fake-licensed license plate identification method provided by the application.

The Memory 111 may be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Read-Only Memory (EPROM), an electrically Erasable Read-Only Memory (EEPROM), and the like.

The processor 112 may be an integrated circuit chip having signal processing capabilities. The Processor 112 may be a general-purpose Processor including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components.

It will be appreciated that the configuration shown in FIG. 1 is merely illustrative and that the electronic device 110 may include more or fewer components than shown in FIG. 1 or may have a different configuration than shown in FIG. 1. The components shown in fig. 1 may be implemented in hardware, software, or a combination thereof. For example, the electronic device 110 may be a computer, a server, a tablet computer, etc., and thus, the present application is not limited to the specific type of the electronic device 110.

On the basis of the electronic device 110 shown in fig. 1, an embodiment of the present application further provides a license plate fake-license plate identification method, which may be applied to the electronic device 110, please refer to fig. 2, where fig. 2 is a schematic flow diagram of a license plate fake-license plate identification method provided in an embodiment of the present application, and the license plate fake-license plate identification method may include the following steps:

s200, acquiring a snapshot sequence of the target license plate; the snapshot sequence comprises a plurality of bayonet identifications, one bayonet identification corresponds to one snapshot time, and the plurality of bayonet identifications are arranged according to the sequence of the snapshot times.

The target license plate in the embodiment of the application can be understood as a license plate number, the checkpoint identifier can be understood as an identification code representing a checkpoint camera, and the snapshot time is the time for snapshot of the license plate by the checkpoint.

And S210, acquiring all target gate identification groups in the snapshot sequence.

Each target bayonet identification group comprises a plurality of target bayonet identifications which are arranged according to the sequence of the snapshot time, and any two adjacent target bayonet identifications meet the preset condition; and the preset condition is characterized in that the difference value of two snapshot times corresponding to any two adjacent target bayonet identifications is greater than or equal to the shortest passing time between two bayonets corresponding to any two adjacent target bayonet identifications.

It can be understood that, for each target bayonet identification group acquired in the snapshot sequence, the following conditions are satisfied: each target bayonet identification group comprises a plurality of target bayonet identifications which are arranged according to the sequence of the snapshot time, and any two adjacent target bayonet identifications meet the preset condition; and the preset condition is characterized in that the difference value of two snapshot times corresponding to any two adjacent target bayonet identifications is greater than or equal to the shortest passing time between two bayonets corresponding to any two adjacent target bayonet identifications.

If the time difference value of the two checkpoints capturing the target license plate is larger than or equal to the shortest passing time of the vehicles between the two checkpoints, it is reasonable that the capturing time of the target license plate capturing the target license plate at the two checkpoints passes through the two checkpoints. Further, the above "any two adjacent target gate identifiers both satisfy the preset condition" may also be understood as that the vehicle corresponding to the target license plate sequentially passes through a plurality of gates corresponding to a plurality of target gate identifiers in the target gate identifier group according to the order of the snapshot time. Therefore, one target checkpoint identifier group represents one space-time trajectory sequence of the vehicle corresponding to the target license plate.

And S220, determining the number of the fake plate of the target license plate according to the number of the target checkpoint identification groups.

It can be understood that, in a period of time, if the number of the target gate identifier groups exceeds 1, it indicates that a vehicle with a plurality of license plates all being target license plates is running in the period of time, and therefore, the number of the fake license plates of the target license plates can be determined according to the number of the target gate identifier groups.

It should be noted that, the bayonet in the embodiment of the present application refers to not only a bayonet camera, but also an electric alarm camera, a general camera, a surveillance camera, and the like in practical applications. That is, the bayonet in the embodiment of the present application refers to a camera.

For the above method embodiments S200 to S220, it should be understood that each target bayonet identification group includes a plurality of target bayonet identifications arranged according to the order of the snapshot time, and any two adjacent target bayonet identifications all satisfy the preset condition; and the preset condition is characterized in that the difference value of two snapshot times corresponding to any two adjacent target bayonet identifications is greater than or equal to the shortest passing time between two bayonets corresponding to any two adjacent target bayonet identifications. Therefore, a plurality of target checkpoint markers in the target checkpoint marker group actually represent a space-time trajectory sequence of the target license plate. And in a period of time, if the number of the target bayonet identification groups exceeds 1, it is indicated that a plurality of vehicles with target license plates all run in the period of time, therefore, the number of the fake license plates of the target license plates can be determined according to the number of the target bayonet identification groups, so that the number of the fake license plates can be effectively identified, and the accuracy of fake license plate identification of the license plates is improved.

Further, in some possible embodiments, one possible implementation manner of S200 of the license plate fake plate identification method shown in fig. 2 is as follows, please refer to fig. 3, and S200 may include:

s201, acquiring a plurality of license plate data captured by a plurality of checkpoints; one license plate data corresponds to one snapshot time and one checkpoint identifier.

In the embodiment of the present application, the license plate data may include a license plate image of the vehicle.

When a certain gate takes a snapshot of the license plate of a passing vehicle, the passing vehicle may be taken several times by the gate. In order to reduce the data volume and improve the calculation efficiency, the license plate image of the vehicle captured by the gate, the capturing time and the identifier of the gate can be used as license plate data only when the vehicle passes through the gate.

S202, identifying the numbers in the plurality of license plate data to identify a plurality of target license plate data belonging to the target license plate in the plurality of license plate data.

In the embodiment of the application, the plurality of license plate data correspond to a plurality of license plates, and the target license plate data belonging to the target license plate in the plurality of license plate data is identified by identifying the numbers in the plurality of license plate data. The license plate numbers of the obtained plurality of target license plate data are all consistent with the target license plate.

Since the common license plate is usually in the form of: the image recognition method of the existing license plate is low in accuracy rate when recognizing Chinese characters, so that only the digital part in the license plate image can be recognized when recognizing the license plate image of the license plate data, and a specific license plate number can be recognized. Eliminating the influence of Chinese character recognition errors.

S203, determining a snapshot sequence of the target license plate according to the snapshot time and the checkpoint identification corresponding to the plurality of target license plate data.

For example, the form of the captured snapshot sequence of the target license plate may be as follows:

the sequence (i.e. snapshot sequence) of the passing checkpoint marks of the target license plate according to the sequence of the snapshot time is C_seq＝[C₁,C₂,C₃,C₄,C₅,C₆,C₇,C₈]The time stamp (i.e. snapshot time) passing through each gate mark is T_seq＝[‘2020/5/4 20:10:00’,‘2020/5/4 20:12:05’,‘2020/5/4 20:12:45’,‘2020/5/4 20:14:52’,‘2020/5/4 20:17:02’,‘2020/5/4 20:18:11’,‘2020/5/4 20:17:02’,‘2020/5/4 20:18:45’]. Can understand C_seqSequence and T_seqOne-to-one correspondence of elements in a sequence, e.g. bayonet C₁Corresponding to the snap shots '2020/5/420: 10: 00'.

Further, in some possible embodiments, one possible implementation manner of S210 of the license plate fake plate identification method shown in fig. 2 is as follows, please refer to fig. 4, and S210 may include:

s211, acquiring a first-order shortest travel time matrix of a plurality of checkpoints adjacent to the checkpoints; the first-order adjacent bayonet shortest travel time matrix represents the shortest passing time between any two bayonets in the plurality of bayonets.

In this embodiment of the application, a first-order bayonet adjacent matrix of a plurality of bayonets corresponding to a plurality of bayonet identifiers may be obtained first, and then a first-order adjacent bayonet shortest travel time matrix of the plurality of bayonets may be obtained according to the first-order bayonet adjacent matrix (see S211A to S211H below). The first-order bayonet adjacency matrix representation is the adjacent relation between any two bayonets in the plurality of bayonets.

For example, assume that the plurality of bayonet markers includes: c1, C2, C3, C4, C5, C6, C7, and C8, the first-order bayonet adjacency matrix of the bayonets corresponding to the plurality of bayonet markers may be obtained as follows:

wherein the content of the first and second substances,

for example, A₁₂The bayonet corresponding to C1 and the bayonet corresponding to C2 are adjacent bayonets, 1.

Then, a first-order shortest travel time matrix of the adjacent gates can be obtained according to the matrix A and the historical vehicle passing data shot by each gate. For example, the first-order adjacent bayonet shortest travel time matrix of a plurality of bayonets corresponding to a plurality of bayonet identifications may be as follows:

wherein, T_ijThe value of (1) is the shortest passing time (unit can be seconds) of the vehicle passing through the two bayonets, inf is infinite, and the two bayonets are not adjacent to each other, so that the vehicle cannot pass through the two bayonets in sequence.

In the embodiment of the application, most application scenarios can directly utilize the first-order adjacent gate shortest travel time matrix to obtain a reasonable track of the target vehicle. That is, all target gate identifier groups in the snapshot sequence are obtained according to the first-order shortest travel time matrix of the adjacent gates.

The inventor finds that the abnormality of the road network equipment can cause the license plate of the vehicle to be missed by the gate, and the abnormality of the road network equipment mainly comprises the following steps: the condition that certain or some small-range cameras are missed due to the conditions of municipal construction, municipal power failure, damage to front-end camera equipment and the like. For the condition of device missing, two originally non-adjacent gates are put together from the front and back of the vehicle-passing snapshot data, so that the target gate identification group (namely the track of the target vehicle) judged by the first-order adjacent gate shortest travel time matrix is unreasonable. Because of the possibility of missed shots, the otherwise non-adjacent bayonet in the first order bayonet abutment matrix becomes an abutment, but the abutment does not appear in the first order abutment bayonet shortest travel time matrix (or travel time is infinite).

For example, A in matrix A₁₃0, which is not contiguous, resulting in a in matrix a due to missed shots at the bayonet₁₃Becomes 1 and C1 becomes contiguous with C3. However, since the first-order shortest travel time matrix of the adjacent gate is obtained from the historical mass vehicle-passing sequence data through the historical data, the result of the default historical mass data statistics (removing abnormal values) is accurate, and therefore, in the first-order shortest travel time matrix T of the adjacent gate, T₁₃If the value is still inf, it indicates that C1 is not adjacent to C3, and at this time, all target gate mark groups in the snapshot sequence are obtained by using the first-order adjacent gate shortest travel time matrix, and the determined trajectory of the target vehicle is not reasonable. ). Therefore, in order to solve the problem, the above-mentioned method should be usedThe shortest travel time matrix of the adjacent gates is processed to a certain extent (S212 and S213 below).

In addition, in this embodiment of the application, after obtaining the first-order bayonet adjacent matrix of the multiple bayonets, the method may further include: determining a second order bayonet abutment matrix of the plurality of bayonets according to the first order bayonet abutment matrix and the following formula:

wherein A is a first order bayonet adjacent matrix, A^*Being a second order bayonet adjacent matrix, an indication of a matrix multiplication,

and (4) representing addition of corresponding elements of the matrix, wherein trunc is a truncation function, when the element is greater than or equal to 1, 1 is used for truncation, and otherwise, 0 is output.

For example, according to a first order bayonet abutment matrix and the above formula, a second order bayonet abutment matrix for a plurality of bayonets may be determined as follows:

the second-order bayonet adjacency matrix can reflect the second-order adjacency relation in the road network bayonet, can still effectively process certain single cameras under the condition of missing shooting, and increases the robustness of the license plate fake plate identification method provided by the embodiment of the application.

S212, determining a second-order shortest travel time matrix of the plurality of the bayonets according to the first-order shortest travel time matrix of the adjacent bayonets and the following formula:

wherein, T_ikElement of ith row and kth column of the ith shortest travel time matrix of first-order adjacent bayonet, T_kjIs the element of the ith row and the kth row of the first-order shortest travel time matrix adjacent to the bayonet, k is the serial number of the bayonet mark, N is the total number of the bayonet mark,

is the element of the ith row and the jth column of the second-order adjacent bayonet shortest travel time matrix.

For example, according to the first-order adjacency-bayonet shortest travel time matrix and the above formula, the second-order bayonet adjacency matrix of the plurality of bayonets determined may be as follows:

and S213, acquiring all target gate identification groups in the snapshot sequence according to the second-order adjacent gate shortest travel time matrix.

In the process of acquiring all target gate identifier groups in the snapshot sequence according to the second-order shortest travel time matrix of the adjacent gates, it is determined whether a difference between two snapshot times corresponding to two gate identifiers is greater than or equal to a shortest transit time between two gates corresponding to the two gate identifiers (that is, the two gate identifiers are subjected to rationality determination to determine the rationality of the target vehicle passing through the two gates, which may refer to steps S210A-2 and the like). The shortest elapsed time between two bayonets corresponding to two respective bayonet marks is the shortest value of the sum of the elapsed times of all the bayonets constituting the second-order adjacency. For example, C1 and C3 are not contiguous, but C1 and C3 may be second-order contiguous by two sets of intermediate bayonets, e.g., C1 and C4 contiguous, C4 and C3 contiguous; or C1 and C5 are adjacent, and C5 and C3 are adjacent. The shortest time here refers to the minimum value of the total time in the two groups (first group C1, C4, C3, second group C1, C5, C3), and assuming that the time taken by the first group (a1 and a4 are adjacent, a4 and A3 are adjacent) is 60 seconds (a1 to a4 shortest time is 20 seconds, a4 to A3 shortest time is 40 seconds), the time taken by the second group is 80 seconds (a1 to A5 shortest time is 50 seconds, A5 to A3 shortest time is 30 seconds), then the second order adjacent shortest time of a1 to A3 is 60 seconds of the time taken by the first group.

Further, according to the first-order adjacent bayonet shortest travel time matrix and formula

In the process of determining the second-order adjacent bayonet shortest travel time matrix of a plurality of bayonets, each actually determined

Is actually a bayonet C_iTo C_jThe shortest value (min (T) of the sum of the time spent by the card ports in all second-order adjacency relations_ik+T_kj) However, this ignores bayonet C_iAnd C_jThe transit time between the two adjacent to each other, therefore, S212 can be further optimized as follows:

determining a second-order adjacent gate shortest travel time matrix of a plurality of gates according to the first-order adjacent gate shortest travel time matrix and the following formula:

wherein, T_ikElement of ith row and kth column of the ith shortest travel time matrix of first-order adjacent bayonet, T_kjIs the element of the kth row and the ith column of the kth row of the shortest travel time matrix of the first-order adjacent bayonet, N is the total number of the bayonets,

is the element of the ith row and the jth column of the second-order adjacent bayonet shortest travel time matrix. The addition operation and the minimum operation may be performed on an infinite value (inf), and it is agreed that the result of "inf + inf" or "c + inf" is still "inf" (infinity), "min (c, inf)" is c, "min (inf ) is inf", and c is a constant.

It can be understood that since

Fastening opening C_iAnd C_jThe time of passage between the two when the two are adjacent to the bayonet C_iTo C_jThe sum of the time used by the card ports in all the second-order adjacency relations is comprehensively considered, and the minimum value is taken as the minimum value

The value of the second-order adjacent bayonet shortest travel time matrix is more appropriate to the actual situation, the method provided by the embodiment of the application has stronger robustness to the real scene, the method is more scientific, and the accuracy of the obtained result is higher.

It can be understood that inf in the second-order adjacent gate shortest travel time matrix indicates that two corresponding gates are not adjacent (also called non-connected), so that when two adjacent gates in the snapshot sequence are not connected, the two corresponding gates can be directly judged to be unreasonable. Therefore, the connectivity and the space-time rationality of the vehicle running through the bayonet can be judged only by the second-order adjacent bayonet shortest travel time matrix T.

It should be understood that the missed-shoot of the gate exists objectively, and it is impossible to determine in advance whether the missed-shoot occurs in the gate, and the missed-shoot may make the trajectory of the target vehicle determined by the first-order adjacent gate shortest travel time matrix unreasonable. According to the method provided by the embodiment of the application, the problem caused by missed shooting can be solved by constructing the second-order adjacent bayonet shortest travel time matrix, so that the method for identifying the fake plate of the license plate provided by the embodiment of the application has stronger robustness on a real scene, is more scientific, and has higher accuracy of the obtained result.

Further, in some possible embodiments, referring to fig. 5, fig. 5 provides a possible implementation manner of the license plate fake plate identification method S211 shown in fig. 4, and S211 may include:

S211A, acquiring a first number of times that a plurality of vehicles pass through a first target gate within a preset time period.

For example, the number of all vehicles passing through the first target gate in the past 5 days is acquired as the first count.

And S211B, acquiring a second frequency of the multiple vehicles continuously passing through the first target gate and the second target gate in the preset time period.

The vehicle continuously passes through the first target bayonet and the second target bayonet, namely, the vehicle passes through the second target bayonet immediately after passing through the first target bayonet.

For example, the number of all vehicles passing through the first target gate and the second target gate in succession in the same past 5 days is acquired as the second count.

S211C, a ratio of the second number to the first number is obtained.

Theoretically, the first target gate and the second target gate which are not adjacent to each other cannot allow the vehicle to pass through continuously, that is, the second time is 0. However, in practical applications, the second number may be greater than 0, since data may be erroneous. However, the probability of data errors is often small, and therefore the ratio of the second number to the first number is small.

Also, it is understood that the ratio of the second number of times to the first number of times characterizes the likelihood that the first target bayonet is adjacent to the second target bayonet.

S211D, judging whether the ratio is larger than a preset threshold value; if so, determining that the first target bayonet and the second target bayonet are adjacent bayonets; if not, determining that the first target bayonet and the second target bayonet are not adjacent bayonets.

The preset threshold may be set according to road network characteristics (the value may be an empirical value), such as 1%, 5%, 10%, etc., which is not limited in this application.

And S211E, repeating the steps until the adjacent relation between any two bayonets in the plurality of bayonets is obtained.

S211F, the shortest elapsed time of the vehicle between any two adjacent gates among the plurality of gates is obtained.

S211G, determining a first-order bayonet adjacent matrix of the plurality of bayonets according to the adjacent relation between any two bayonets in the plurality of bayonets.

And S211H, determining the shortest travel time matrix of the first-order adjacent gates of the plurality of gates according to the shortest vehicle passing time between the first-order adjacent gates and any two adjacent gates of the plurality of gates.

The above-mentioned S211A to S211H will be further explained in connection with practical applications.

Firstly, obtaining all passing data of all gates of a road network captured for a period of time from a passing database, grouping license plates, and forming a gate sequence set Cset according to the sequence of the license plates passing the gates.

Specifically, assume that there are m license plates in the passing database, and that the license plate i (i ═ 1,2, …, m) passes n_iA bayonet is respectively

Bayonet sequence set C formed by M license plates_set；

Wherein, C_setEach element of (a) represents a bayonet sequence.

Then, all license plates pass through the bayonet sequence to be split into bayonet pairs (C)_up,C_dowm) One pair represents one pass of the vehicle from the upstream gate to the downstream gate. Specifically, for a bayonet sequence passed by each license plate, a sliding window is carried out by taking 2 as the window size to take out bayonet pairs, and the bayonet pairs slide from a first bayonet to a last bayonet to form n-1(n represents the number of bayonets passed by the license plate) bayonet pairs.

Then, taking the bayonet pairs split from all license plates as basic data, and counting the vehicles from the upstream bayonet C on the basis of each upstream bayonet_upTo respective downstream bayonet C_dowmNumber of vehicle passing occurrences n_ijAnd calculating it at all C_iRatio r in the passing record of upstream card port_ij。

Wherein, the formula is: n is_ij＝∑(c_up＝c_i,c_down＝c_cj),r_ij＝n_ij/∑(c_up＝c_i)。

Setting a threshold value theta according to the road network characteristics, if r_ijIf the value is larger than theta, the bayonet pair is considered as an adjacent bayonet, otherwise, the upstream bayonet and the downstream bayonet are not considered as adjacent. In this way, a bayonet abutment matrix a is constructed for all bayonets. Where A ∈ R^N×N(N represents the number of gates in the road network), and

then, for all adjacent bayonets C_iAnd C_jAnalyzing the passing time distribution according to the passing record, providing an abnormal value according to an abnormal detection algorithm, judging the normal time distribution range, and finding the shortest passing time Tmin of the gate_ijAnd forming a first-order shortest travel time matrix T adjacent to the bayonet. T is_ijAnd A_ijAnd correspond to each other. In particular, when A_ijWhen 0, T_ijIs infinite.

Specifically, there are many abnormality detection algorithms, and the normal distribution of time can be calculated from the IQR value, and if the passing time data of 1.5 times IQR less than the first quartile is regarded as an abnormal passing record and is cut off by the value as the minimum passing time Tmin of the gate_ij(ii) a It is also possible to find the minimum transit time Tmin by the 3 σ rule assuming that the passing time distribution between the adjacent gates follows a gaussian distribution_ij。

At the shortest time Tmin_ijAnd the judgment result is used as a standard and is used as a judgment basis for space-time rationality, and if the travel time of a certain license plate at an adjacent gate is found to be less than the minimum time in the judgment process, the judgment is abnormal.

It should be understood that in the embodiment of the application, a large amount of historical vehicle passing data in a road network is utilized, and the shortest time of the adjacent gates is taken as a standard, so that the main judgment means of the connectivity and the space-time rationality of the two adjacent gates in the snapshot sequence is adopted. Namely, the road network vehicle passing data is used for calculating the space-time rationality and the checkpoint connectivity judgment standard value, so that the license plate fake plate identification method provided by the embodiment of the application has higher result reliability.

In this embodiment of the application, a first mount identifier of the snapshot sequence may be preset in the electronic device 110 as a current mount identifier, and a reasonable track list, an abnormal mount list, and a suspended mount list are preset, where the reasonable track list is used to store the target mount identifier group, and the suspended mount list is used to store the suspended mount identifier.

Further, the embodiment of the present application also provides a possible implementation manner of S210 of the method for recognizing a license plate in a fake plate as shown in fig. 2, and for more clearly describing the embodiment, the following assumptions are made before:

the electronic device 100 pre-stores a reasonable trajectory list palst (used for storing a target checkpoint identifier group, which is an empty subsequence list during initialization and represents a continuous space-time trajectory of a license plate) in a storage medium; an abnormal bayonet list outlst (used for storing an abnormal bayonet identification and indicating that a target license plate captured by a corresponding bayonet is abnormal); a list hangtab hanglist hanglst (for storing hangtab identification, representing the tab to be checked).

Taking the snapshot sequence of a certain license plate p (target license plate) as an example, it is assumed that the snapshot sequence of a certain license plate p passing through within the statistical time range is C_seqSequence C thereof_seq＝[C₁,C₂,C₃,…,C_n]The corresponding snapshot time of each bayonet capturing the license plate p is T_seq＝[t₁,t₂,t₃,…,t_n]。

First, referring to fig. 6, S210 (step of acquiring all target mount id groups in the snapshot sequence) may include:

S210A-1, acquiring a current bayonet identification and a first bayonet identification adjacent to the current bayonet identification.

For example, first, the sequence C is acquired_seqC in (1)₁And C₂，C₁Is marked by the current bayonet C₂And the first bayonet identification adjacent to the current bayonet identification is obtained by analogy, and when the step is returned every time, two adjacent bayonet identifications are obtained.

S210A-2, judging whether the difference value between the two capturing times respectively corresponding to the current bayonet identification and the first bayonet identification is larger than or equal to the shortest passing time between the two bayonets respectively corresponding to the current bayonet identification and the first bayonet identification.

The difference value of two snapshot times corresponding to the current gate identifier and the first gate identifier respectively represents the difference between the time when the two gates corresponding to the two gate identifiers snapshot the target license plate; the shortest passing time between two bayonets corresponding to the current bayonet identification and the first bayonet identification respectively represents the shortest passing time of a vehicle between the two bayonets corresponding to the two bayonet identifications.

S210A-3, if yes, copying the current bayonet identification and the first bayonet identification to a target bayonet identification group to which the current bayonet identification belongs.

If the difference value of the two snapshot times respectively corresponding to the current bayonet identification and the first bayonet identification is larger than or equal to the shortest passing time between the two bayonets respectively corresponding to the current bayonet identification and the first bayonet identification, it is reasonable that the target license plate passes through the two bayonets, and then the two bayonet identifications can together represent a section of space-time trajectory of the target license plate. Furthermore, the current gate identifier and the first gate identifier may be copied to a target gate identifier group to which the current gate identifier belongs, so as to determine a space-time trajectory sequence of the vehicle corresponding to the target license plate.

S210A-4, if not, executing a first hanging bayonet judging step.

If the difference value of the two capturing times respectively corresponding to the current bayonet identification and the first bayonet identification is not larger than or not equal to the shortest passing time between the two bayonets respectively corresponding to the current bayonet identification and the first bayonet identification, it is unreasonable to indicate that the target license plate passes through the two bayonets, and then the two bayonet identifications cannot indicate a section of space-time trajectory of the target license plate, and a first hanging bayonet judgment step needs to be executed for further judgment.

S210A-5, taking the first bayonet identification as the current bayonet identification, and repeatedly executing the steps until the snapshot sequence is traversed, so as to obtain all target bayonet identification groups in the snapshot sequence.

E.g. C₂As the first bayonet sign, firstly C₂And as the current bayonet identification, returning to execute S210A-1 until the snapshot sequence is traversed, and obtaining all target bayonet identification groups in the snapshot sequence.

It should be noted that, in the description in the embodiment of the present application, it is reasonable that one bayonet is reasonable from another bayonet, which means that it is reasonable that the target license plate passes through the two bayonets, and then the two bayonet identifications can together represent a segment of space-time trajectory of the target license plate.

The above-mentioned S210A-1 to S210A-5 are further explained below in connection with practical applications.

First, C is initially selected₁As the current bayonet sign C_currThe variable i is initialized and set to 2.

Extracting element No. i C from snapshot sequence_i(i is less than or equal to n) as the first bayonet mark adjacent to the current bayonet mark, and judging C₁To C_iThe spatiotemporal rationality of (i.e., execution of S210A-2).

If C₁To C_iIf the judgment result of the space-time rationality is yes, C is added_currAnd C_iAdd to its preamble bayonet (last one to C)_currSpace-time reasonable bayonet) is located, if C_currIf there is no preamble, add 1 reasonable sequence track list in the tralst and add C_currAnd C_iIs added thereto and is mixed with C_iAs the current bayonet C_currLet i be i +1, the process returns to S210A-1.

If C₁To C_iIf the result of the spatio-temporal rationality judgment is no, S210B is executed.

Further, referring to fig. 7, the step of executing the first hanging bayonet determination S210B may include:

S210B-1, judging whether a first target hanging bayonet mark exists in the hanging bayonet list or not; the difference value of the two capturing times respectively corresponding to the first target hanging bayonet identification and the first bayonet identification is larger than or equal to the shortest passing time between the two bayonets respectively corresponding to the first target hanging bayonet identification and the first bayonet identification.

S210B-2, if yes, copying the first mount identifier to a target mount identifier group to which the first target suspension mount identifier belongs, and deleting the first target suspension mount; copying the current bayonet identification as a hanging bayonet identification to a hanging bayonet list; and taking the first bayonet identification as the current bayonet identification, and returning to execute the step of obtaining the current bayonet identification and the first bayonet identification adjacent to the current bayonet identification.

It can be understood that if the first target hanging bayonet mark exists in the hanging bayonet list, it indicates that the bayonet corresponding to the first target hanging bayonet mark is reasonable to the bayonet corresponding to the first bayonet mark (it is reasonable that the target license plate passes through the two bayonets).

S210B-3, if not, and the snapshot sequence is not traversed completely, executing an abnormal bayonet judgment step.

It can be understood that if the first target hanging bayonet identification does not exist in the hanging bayonet list, it indicates that a reasonable hanging bayonet identification corresponding to the first bayonet identification does not exist in the hanging bayonet list.

The above-mentioned S210B-1 to S210B-3 are further explained below in connection with practical applications.

Sequentially scanning each hanging bayonet identification in the hanging bayonet list from back to front, and checking whether a hanging bayonet identification exists or not to check whether a hanging bayonet identification exists from C_iIt is reasonable (i.e. implement S210B-1) if a certain suspension bayonet C_mTo C_iSpace-time is reasonable, then C_mIs added to C_mThe target card gate where the preorder card gate is located marks the end of the group, and marks C_mDeleting C from the hanging bayonet list_currAdd to the end of the hang-up bayonet list and add C_iAs the current bayonet C_currLet i be i +1, the process returns to S210A-1.

If all the suspension bayonet identifiers in the suspension bayonet list are C_iAll are not reasonable, then judge C_iWhether or not to reach C_seqAt the end of the sequence, if C_iHas arrived at C_seqAt last, C is added_iAnd adding the license plate number to the abnormal checkpoint list outlst, and determining that the fake plate identification method of the license plate is completed. If C_iNot reaching C_seqAt the end, the abnormal bayonet determination step S210C is executed.

Further, referring to fig. 8, the step of executing the abnormal bayonet determination S210C may include:

S210C-1, acquiring a second bayonet sign adjacent to the first bayonet sign.

It will be appreciated that in the snapshot sequence C_seq，C_iIs firstBayonet identification, then C_i+1Is a second bayonet sign adjacent to the first bayonet sign.

S210C-2, judging whether the difference value of the two capturing times respectively corresponding to the current bayonet identification and the second bayonet identification is larger than or equal to the shortest passing time between the two bayonets respectively corresponding to the current bayonet identification and the second bayonet identification.

S210C-3, if yes, copying the second bayonet identification to a target bayonet identification group to which the current bayonet identification belongs, and copying the first bayonet identification to an abnormal bayonet list; and taking the second bayonet identification as the current bayonet identification, and returning to execute the step of acquiring the current bayonet identification and the first bayonet identification adjacent to the current bayonet identification.

It can be understood that if the difference between the two capturing times respectively corresponding to the current gate identifier and the second gate identifier is greater than or equal to the shortest elapsed time between the two gates respectively corresponding to the current gate identifier and the second gate identifier; the bayonet corresponding to the current bayonet identification is indicated, and the bayonet corresponding to the second bayonet identification is reasonable (the target license plate is reasonable when passing through the two bayonets).

S210C-4, if not, executing a second hanging bayonet judging step.

It can be understood that if not, it is unreasonable to say that the bayonet corresponding to the current bayonet identification is connected to the bayonet corresponding to the second bayonet identification.

The above-mentioned S210C-1 to S210C-4 will be further explained with reference to practical applications.

First, check the bayonet C_currTo C_i+1Whether it is reasonable (i.e., execution of S210C-2).

If the execution result of S210C-2 is YES, then C is added_iAdd to Exception Bayonet List outlst (Add to List Final), C_currAnd adding the identifier to the tail of the target bayonet identifier group where the preamble bayonet is positioned. C is to be_i+1As the current bayonet C_currLet i be i +2, the process returns to S210A-1.

If the execution result of S210C-2 is no, the second hanging bayonet determining step S210D is executed.

Further, referring to fig. 9, the step of executing the second hanging bayonet determination S210D may include:

S210D-1, judging whether a second target hanging bayonet mark exists in the hanging bayonet list or not; the difference value of the two capturing times respectively corresponding to the second target hanging bayonet identification and the second bayonet identification is larger than or equal to the shortest passing time between the two bayonets respectively corresponding to the second target hanging bayonet identification and the second bayonet identification.

S210D-2, if yes, copying the second mount identifier to a target mount identifier group to which the second target mount belongs, and deleting the second target mount; copying the first bayonet identification to an abnormal bayonet list; and taking the second bayonet identification as the current bayonet identification, and returning to execute the step of acquiring the current bayonet identification and the first bayonet identification adjacent to the current bayonet identification.

It can be understood that if a second target hanging bayonet mark exists in the hanging bayonet list, it is indicated that the bayonet corresponding to the second target hanging bayonet mark is reasonable to the bayonet corresponding to the second bayonet mark (it is reasonable that the target license plate passes through the two bayonets).

S210D-3, if not, executing a second track identification step.

It can be understood that if the card identifier corresponding to the second card identifier is not in the card identifier list, the card identifier corresponding to the second card identifier is not in the card identifier list.

The above-mentioned S210D-1 to S210D-3 are further explained below in connection with practical applications.

Sequentially scanning each hanging bayonet identification in the hanging bayonet list from back to front, and checking whether a hanging bayonet to C exists_i+1If it is reasonable (i.e., execution S210D-1), if a certain suspension bayonet C_mTo C_i+1Space-time is reasonable, then C_mIs added to C_mThe target card gate where the preorder card gate is located marks the end of the group, and marks C_mDeleting C from the hanging bayonet list_iAdd to the abnormal bayonet list outlst, add C_currAdd to the end of the hang-up bayonet list and add C_i+1As the current bayonet C_currLet i be i +2, the process returns to S210A-1. If it is notAll suspension bayonets to C_i+1Are all unreasonable, then pair C_iTo C_i+1The determination of the spatiotemporal rationality corresponds to the execution of the second trajectory identification step S210E.

Further, referring to fig. 10, the executing the second track identifying step S210E may include:

S210E-1, judging whether the difference value of the two capturing times respectively corresponding to the first bayonet identification and the second bayonet identification is larger than or equal to the shortest passing time between the two bayonets respectively corresponding to the first bayonet identification and the second bayonet identification.

S210E-2, if yes, adding a target bayonet identification group in the reasonable track list, and copying the first bayonet identification and the second bayonet identification to the added target bayonet identification group; and taking the second bayonet identification as the current bayonet identification, and returning to execute the step of acquiring the current bayonet identification and the first bayonet identification adjacent to the current bayonet identification.

It can be understood that if the difference between the two capturing times respectively corresponding to the first bayonet identification and the second bayonet identification is greater than or equal to the shortest elapsed time between the two bayonets respectively corresponding to the first bayonet identification and the second bayonet identification; it is reasonable that the target license plate passes through the two checkpoints, and the two checkpoint identifications can represent a section of space-time trajectory of the target license plate together.

S210E-3, if not, copying the first bayonet identification to an abnormal bayonet list, taking the second bayonet identification as the current bayonet identification, and returning to execute the step of obtaining the current bayonet identification and the first bayonet identification adjacent to the current bayonet identification.

The above-mentioned S210E-1 to S210E-2 are further explained below in connection with practical applications.

If C_iTo C_i+1If the time-space rationality judgment result is yes, C is used_iForming a new driving track in the reasonable track list tralst for the starting bayonet mark, namely adding a target bayonet mark group in the track list tralst, and adding C_iAdding into the target bayonet identification group with C_i+1For the current bayonet sign C_currLet us orderi +2, returning to execute S210A-1;

if C_iTo C_i+1If the time-space rationality judgment result is negative, determining the bayonet identification C_iTo bayonet sign C_i+1If not, C is added_iAdd to the abnormal bayonet list, let i ═ i +2, return to execution S210A-1.

During the above operation, if the first bayonet sign has reached C_seqAnd at the end, adding all the bayonets which are unreasonably detected into an abnormal bayonets list, and adding all the suspended bayonets into a track list where the pre-positioned bayonets are positioned.

And judging the number N (sub-list number) of the target card port identification groups in the reasonable track list tralst, if N is greater than 1, determining that the fake plate exists, and determining that the number of fake plate vehicles is N-1, otherwise, determining that the fake plate does not exist.

It can be understood that the embodiment of the application can generate the adjacency relation by passing a large amount of vehicle-passing data through the bayonet, judge the adjacency of two bayonets (direct adjacency and second-order adjacency), and calculate the minimum travel time required between adjacent bayonets by passing a large amount of vehicle-passing data through the road network, and the data is more reasonable and scientific. Originally, the space-time rationality of a vehicle passing through two bayonets is judged by calculating the longitude and latitude distances of the two bayonets so as to calculate the average speed. The existing method cannot really reflect the direct or indirect connectivity of the two bayonets because the distance calculation error through the longitude and latitude is too large.

To illustrate the application of the embodiments of the present application to a license plate fake-license plate identification method in detail, two examples are specifically given as an illustration, each example represents a different application scenario (whether there is a fake-license plate car and a license plate identification error point), and it should be assumed that the following embodiments are illustrative of the present application and not limiting the use of the present application.

Example 1, a license plate is erroneously identified as the license plate in the license plate trajectory sequence, and no license plate is left.

Fig. 11 is a schematic view of a situation that a road network of a license plate passes through a gate according to an embodiment of the present application. In the passing of the license plate through the checkpoint sequence, no fake license plate (multiple tracks) exists, but the license plate which is wrongly identified exists.

In order to explain the implementation process of the license plate fake plate identification method in the embodiment of the application, the construction process of the second-order bayonet adjacent matrix and the second-order bayonet shortest travel time matrix is not described in detail, and it is assumed that the second-order bayonet adjacent matrix and the second-order bayonet shortest travel time matrix are already constructed.

Suppose that the time-based passage bayonet sequence of a certain license plate is C_seq＝[C₁,C₂,C₃,C₄,C₅,C₆,C₇,C₈]Time stamp T of each card gate_seq＝[‘2020/5/4 20:10:00’,‘2020/5/4 20:12:05’,‘2020/5/4 20:12:45’,‘2020/5/4 20:14:52’,‘2020/5/4 20:17:02’,‘2020/5/4 20:18:11’,‘2020/5/4 20:17:02’,‘2020/5/4 20:18:45’]The second-order adjacent relation of each bayonet is as follows:

the corresponding second order bayonet travel time matrix (in seconds) is:

the following describes the process of the license plate fake-license plate identification method in detail according to the steps:

step 1: various parameters are initialized.

The list of initialized reasonable trajectories, false, is a list of [ 2 ] containing an empty list]](ii) a The abnormal bayonet list outlst is an empty list [ 2 ]](ii) a Bayonet connectivity and space-time rationality judgment condition cond (here, the shortest travel time matrix T of the second-order adjacent Bayonets)^**) (ii) a The suspension bayonet list hanglist is a null list]。

Step 2: grouping all license plates captured within a period of time, sequencing each license plate according to time through a checkpoint to form a checkpoint sequence, and deleting repeated shooting (analyzing) of license plates. Here is C in the above embodiment_seqAnd T_seq。

And step 3: initial selection C₁As the current bayonet C_currThe initialization variable i is 2.

And 4, step 4: taking element No. i C out of bayonet sequence_i(i is less than or equal to n) as the bayonet to be detected, and judging that the current bayonet reaches the bayonet C to be detected_iThe space-time rationality of (a).

Here, the second bayonet C is selected₂As the checkpoint to be checked, according to the second-order adjacent checkpoint shortest travel time matrix T^**Judging the bayonet C₁To bayonet C₂The space-time rationality of (A) is analyzed and found to be that the license plate is from C₁To bayonet C₂The time duration is 125 seconds, the shortest travel time of the two checkpoints is 100 seconds, and the inspection result is reasonable. Fastening opening C₁Added to the list of legitimate bayonets (at this time the list of legitimate bayonets tralst becomes [ [ C ]₁]]) The current bayonet is C₂The bayonet to be inspected is C₃. This step is repeated.

And 5: after inspection, the clamping opening is found to be C₄The travel time is reasonable.

At this time, the reasonable bayonet list tralst becomes [ [ C ]₁,C₂，C₃]]The current bayonet is C₄The bayonet to be inspected is C₅Checking and finding out bayonet C₄To the bayonet is C₅Not reasonable (C)₄To the bayonet is C₅The minimum time is infinite).

Step 6: with bayonet C₆As the bayonet to be inspected, the current bayonet C is found₄To C₆Space-time is reasonable, then the bayonet C is inserted₅Add to abnormal Bayonet List, C₄Add to its preamble bayonet C₃In the list of locations (in this case the list of reasonable checkpoints is [ [ C ]₁,C₂，C₃]])，C₆As current bayonet, C₇For checking the bayonet C₆To bayonet C₇The rationality of (2).

And 7: inspection result issuing bayonet C₆To bayonet C₇The space-time rationality is unreasonable (the minimum travel time between two bayonets is infinite), and C₈As the bayonet to be checked, repeating the step 6, checkingBayonet C₆To C₈The space time of (A) is reasonable.

And 8: the inspection result shows that the bayonet C₆To C₈The space and time of (A) is reasonable, then the bayonet C is inserted₇Adding to the abnormal bayonet list (the abnormal bayonet list outlst is [ C ] at the moment₅,C₇]) Mixing C with₆Add to its preamble bayonet C₄In the track sequence (in this case the list of reasonable checkpoints is [ [ C ]₁,C₂，C₃，C₄，C₆]])，C₈As the current bayonet.

And step 9: examination of hair loss C₈When the current card port C reaches the end of the large card port list₈Add to its preamble bayonet C₆In the track sequence (in this case the list of reasonable checkpoints is [ [ C ]₁,C₂，C₃，C₄，C₆，C₈]])。

Step 10: and checking the number of the tracks in the track sequence, and determining whether the fake plate exists or not and the fake plate times.

Only one track exists in the reasonable track list, and no fake plate exists. Two abnormal checkpoints exist in the abnormal checkpoint list, and two checkpoints with wrong license plate recognition exist in the license plate sequence.

Example 2, a license plate is wrongly identified and a fake license plate is present in the license plate track sequence.

Fig. 12 is a schematic view illustrating a situation where a road network of a license plate passes through a gate according to an embodiment of the present application. In the passing of the license plate through the checkpoint sequence, the condition of fake license plate (a plurality of tracks) exists, and the license plate which is wrongly identified exists.

Suppose that a license plate is C in time sequence through the bayonet_seq＝[C₁,C₂,C₃,C₄,C₅,C₆,C₇,C₈,C₉,C₁₀,C₁₁]Time stamp T of each card gate_seq＝[‘2020/5/4 20:10:00’,‘2020/5/4 20:12:05’,‘2020/5/4 20:12:45’,‘2020/5/4 20:14:52’,‘2020/5/4 20:17:02’,‘2020/5/4 20:18:11’,‘2020/5/4 20:17:02’,‘2020/5/4 20:18:45’,‘2020/5/4 20:31:02’,‘2020/5/4 20:50:02’]The second-order adjacent relation of each bayonet is as follows:

the corresponding second order bayonet travel time matrix (in seconds) is:

step 1: various parameters are initialized.

Step 2: grouping all license plates captured within a period of time, sequencing each license plate according to time through a checkpoint to form a checkpoint sequence, and deleting repeated shooting (analyzing) of license plates. Here, it is the above-mentioned C_seqAnd T_seq。

Here, the second bayonet C is selected₂As a checkpoint to be checked, the travel minimum time matrix T is communicated according to a second-order checkpoint^**Judging the bayonet C₁To bayonet C₂The space-time rationality of (A) is analyzed and found to be that the license plate is from C₁To bayonet C₂The time duration is 125 seconds, the shortest travel time of the two checkpoints is 100 seconds, and the inspection result is reasonable. Fastening opening C₁Add to reasonable bayonet list (at this point)Reasonable bayonet list tralst becomes [ [ C ]₁]]) The current bayonet is C₂The bayonet to be inspected is C₃. This step is repeated.

And 5: after inspection, the clamping opening is found to be C₃The travel time is reasonable.

At this time, the reasonable bayonet list tralst becomes [ [ C ]₁,C₂]]The current bayonet is C₃The bayonet to be inspected is C₄Checking and finding out bayonet C₃To the bayonet is C₄Unreasonable (bayonet C)₃To the bayonet is C₄The minimum time is infinite).

Step 6: with bayonet C₅As the bayonet to be inspected, the current bayonet C is found by inspection₃To C₅Unreasonable space-time (bayonet C)₃To the bayonet is C₄The minimum time is infinite), and the list of the suspension bayonet is empty, so that the space-time rationality judgment is carried out on the non-suspension bayonet and the bayonet to be detected. C is to be₃Add to hang bayonet list (add to hang bayonet list as [ C₃]) Then use bayonet C₄As the current mount, inspection mount C₄To bayonet C₅Space-time rationality.

And 7: inspection finds that the bayonet C₄To bayonet C₅Space-time reasonableness (the used time is more than the minimum time), the bayonet C is put into use₄Added as a new track start point to the second sub-list in the reasonable track list (in this case the reasonable bayonet list is [ [ C ])₁,C₂],[C₄]])，C₅As current bayonet, C₆As the bayonet to be inspected, inspection bayonet C₅To bayonet C₆Space-time rationality.

And 8: checking and finding bayonet C₅To bayonet C₆Space-time is reasonable, then the bayonet C is inserted₅Add to preamble bayonet C₄In the list of places (reasonable bayonet list after adding [ [ C ]₁,C₂],[C₄,C₅]]) By bayonet C₆As current bayonet, C₇As the bayonet to be inspected, inspection bayonet C₆To bayonet C₇The space-time rationality of (a).

And step 9: inspection discovery cardMouth C₆To bayonet C₇If the space-time is not reasonable, C is₆Adding the current suspension bayonet list into a suspension bayonet list, and scanning the suspension bayonet list in sequence, wherein the current suspension bayonet list is [ C ]₃,C₆]Then use C₃As the current mount, inspection mount C₃To bayonet C₇Space-time rationality.

Step 10: inspection found C₃To bayonet C₇Space-time unreasonable, i.e. all suspension bayonets to C₇All unreasonable, begin to check the current bayonet C₃To bayonet C₈Space-time rationality.

Step 11: inspection finds that the current bayonet C₃To bayonet C₈Space-time is reasonable, then C₃Add to its preamble bayonet C₂In the track list (in this case, the reasonable track list is [ [ C ]₁,C₂，C₃],[C₄,C₅]]) And delete it from the hanging bayonet list (at this time the hanging bayonet list is [ C ]₆]) Mixing C with₇Add to Exception Bayonet List (Add post Exception Bayonet List is [ C ]₇]) Mixing C with₈As the current mount, inspection mount C₈To bayonet C₉Space-time rationality.

Step 12: inspection finds that the current bayonet C₈To bayonet C₉Space-time is reasonable, then the bayonet C is inserted₈Add to its preamble bayonet C₃Location list (reasonable track list after addition is [ [ C ]₁,C₂，C₃，C₈],[C₄,C₅]]) With C₉As current bayonet, C₁₀As a mount to be inspected, inspection mount C₉To bayonet C₁₀Space-time rationality.

Step 13: inspection finds that the current bayonet C₉To bayonet C₁₀Unreasonable space-time, adding it to the list of suspension bayonets, scanning the list of suspension bayonets, and only setting the suspension bayonets to [ C ] at the moment₆，C₉]Checking the bayonet C₆To bayonet C₁₀Space-time rationality.

Step 14: inspection finds that the current bayonet C₆To bayonet C₁₀Space-time is reasonable, thenBayonet C₆Add to its preamble bayonet C₅The track list (here the reasonable track list is [ [ C ]₁,C₂，C₃，C₈],[C₄,C₅，C₆]]) And delete it from the list of hanging bayonets (the hanging bayonets are simply [ C ] after deletion₉]) With C₁₀As the current mount, inspection mount C₁₀To bayonet C₁₁Space-time rationality.

Step 15: inspection finds that the current bayonet C₁₀To bayonet C₁₁The space-time is not reasonable, then the bayonet C is turned₁₀Add to hang bayonet list (add to hang bayonet list as [ C₉，C₁₀]) With C₉For the current bayonet, check it to bayonet C₁₁The space-time rationality of (a).

Step 16: inspection finds that the current bayonet C₉To bayonet C₁₁Space-time is reasonable, then the bayonet C is inserted₉Add to its preamble bayonet C₈In the track list, since C is now₁₁Having reached the end of the bayonet sequence, the bayonet C will be suspended₁₀Add to its preamble bayonet C₆In the track sequence, C₁₁Add to its preamble bayonet C₉In the trajectory list, the reasonable estimation list is [ C ] at this time₁,C₂，C₃，C₈，C₉，C₁₁],[C₄,C₅，C₆，C₁₀]]And the check is finished.

And step 17: the analysis finds that the abnormal bayonet is C₇The wrong points can be identified for the license plate; the reasonable estimation list has two reasonable tracks, which indicate that the license plate is suited.

In order to execute the corresponding steps in the above embodiments and various possible manners, another implementation manner of the fake plate identification device for a license plate is given below, please refer to fig. 13, and fig. 13 shows a functional module diagram of the fake plate identification device for a license plate provided by the embodiment of the present application. It should be noted that the basic principle and the generated technical effects of the fake plate identification apparatus 300 for license plates provided in the present embodiment are the same as those of the above embodiments, and for the sake of brief description, no part of the present embodiment is mentioned, and reference may be made to the corresponding contents in the above embodiments. This fake-license plate recognition device 300 of license plate includes: an acquisition module 310 and an identification module 320.

Alternatively, the modules may be stored in a memory in the form of software or Firmware (Firmware) or be fixed in an Operating System (OS) of the electronic device 110 provided herein, and may be executed by a processor in the electronic device 110. Meanwhile, data, codes of programs, and the like required to execute the above modules may be stored in the memory.

The acquisition module 310 is used for acquiring a snapshot sequence of a target license plate; the snapshot sequence comprises a plurality of bayonet identifications, one bayonet identification corresponds to one snapshot time, and the plurality of bayonet identifications are arranged according to the sequence of the snapshot times;

the recognition module 320 is used for acquiring all target gate identifier groups in the snapshot sequence;

each target bayonet identification group comprises a plurality of target bayonet identifications which are arranged according to the sequence of the snapshot time, and any two adjacent target bayonet identifications meet the preset condition; the preset condition is characterized in that the difference value of two snapshot times corresponding to any two adjacent target gate identifiers is greater than or equal to the shortest elapsed time between two gates corresponding to any two adjacent target gate identifiers;

the recognition module 320 is further configured to determine the number of fake plate sets of the target license plate according to the number of the target checkpoint identification groups.

It is to be appreciated that the acquisition module 310 can be utilized to support the electronic device 110 in performing the above-described S200, etc., and/or other processes for the techniques described herein, e.g., S201-S203; the identification module 320 may be used to support the electronic device 110 in performing the above-described S210, S220, etc., and/or other processes for the techniques described herein, e.g., S211 to S214, S211A to S211H, S210A-1 to S210A-5, S210B-1 to S210B-3, S210C-1 to S210C-4, S210D-1 to S210D-3, S210E-1 to S210E-3.

Based on the above method embodiment, the present application further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the license plate fake plate identification method are executed.

Specifically, the storage medium may be a general storage medium, such as a mobile disk, a hard disk, and the like, and when a computer program on the storage medium is executed, the fake plate identification method for the license plate can be executed, so that the problems that the number of fake plates cannot be identified and the identification effect is poor in the existing fake plate identification method when fake plates are identified are solved, and the purposes of effectively identifying the number of fake plates and improving the fake plate identification accuracy of the license plate are achieved.

In summary, the embodiment of the application provides a fake plate identification method and device for a license plate, a storage medium and an electronic device. The method comprises the following steps: acquiring a snapshot sequence of a target license plate; the snapshot sequence comprises a plurality of bayonet identifications, one bayonet identification corresponds to one snapshot time, and the plurality of bayonet identifications are arranged according to the sequence of the snapshot times; acquiring all target gate identification groups in a snapshot sequence; each target bayonet identification group comprises a plurality of target bayonet identifications which are arranged according to the sequence of the snapshot time, and any two adjacent target bayonet identifications meet the preset condition; the preset condition is characterized in that the difference value of two snapshot times corresponding to any two adjacent target gate identifiers is greater than or equal to the shortest elapsed time between two gates corresponding to any two adjacent target gate identifiers; and determining the number of the fake plate of the target license plate according to the number of the target checkpoint identification groups.

Each target bayonet identification group comprises a plurality of target bayonet identifications which are arranged according to the sequence of the snapshot time, and any two adjacent target bayonet identifications meet the preset condition; and the preset condition is characterized in that the difference value of two snapshot times corresponding to any two adjacent target bayonet identifications is greater than or equal to the shortest passing time between two bayonets corresponding to any two adjacent target bayonet identifications. Therefore, a plurality of target checkpoint markers in the target checkpoint marker group actually represent a space-time trajectory sequence of the target license plate. And in a period of time, if the number of the target bayonet identification groups exceeds 1, it is indicated that a plurality of vehicles with target license plates all run in the period of time, therefore, the number of the fake license plates of the target license plates can be determined according to the number of the target bayonet identification groups, so that the number of the fake license plates can be effectively identified, and the accuracy of fake license plate identification of the license plates is improved.

The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present application should be covered within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A fake plate identification method of a license plate is characterized by comprising the following steps:

acquiring all target gate identification groups in the snapshot sequence;

2. The method of claim 1, wherein the step of obtaining a snapshot sequence of the target license plate comprises:

acquiring a plurality of license plate data captured by a plurality of checkpoints; one license plate data corresponds to one snapshot time and one checkpoint identifier;

identifying numbers in the plurality of license plate data to identify a plurality of target license plate data belonging to the target license plate in the plurality of license plate data;

and determining a snapshot sequence of the target license plate according to the snapshot time and the checkpoint identification corresponding to the plurality of target license plate data.

3. The method of claim 1, wherein the step of acquiring all target mount identifier groups in the snapshot sequence comprises:

acquiring a first-order shortest travel time matrix of a plurality of checkpoints adjacent to the checkpoints; the first-order adjacent bayonet shortest travel time matrix represents the shortest elapsed time between any two bayonets in a plurality of bayonets;

determining a second-order adjacent bayonet shortest travel time matrix of a plurality of the bayonets according to the first-order adjacent bayonet shortest travel time matrix and the following formula:

wherein, T_ikIs the element of the ith row and the kth column of the first-order adjacent bayonet shortest travel time matrix, T_kjIs the element of the ith row and the ith column of the kth row of the first-order adjacent bayonet shortest travel time matrix,

the element of the ith row and the jth column of the second-order adjacent bayonet shortest travel time matrix is shown in the specification;

and acquiring all target gate identification groups in the snapshot sequence according to the second-order adjacent gate shortest travel time matrix.

4. The method of claim 3, wherein the step of obtaining a first order adjacent bayonet shortest travel time matrix for a plurality of said bayonets comprises:

acquiring a first number of times that a plurality of vehicles pass through a first target gate within a preset time period;

acquiring a second number of times that the plurality of vehicles continuously pass through the first target gate and the second target gate within the preset time period;

acquiring the ratio of the second times to the first times;

judging whether the ratio is larger than a preset threshold value or not; if so, determining that the first target bayonet and the second target bayonet are adjacent bayonets; if not, determining that the first target bayonet and the second target bayonet are non-adjacent bayonets;

repeatedly executing the steps until the adjacent relation between any two bayonets in the plurality of bayonets is obtained;

obtaining the shortest vehicle passing time between any two adjacent bayonets from the plurality of bayonets;

determining a first-order bayonet adjacent matrix of the plurality of bayonets according to the adjacent relation between any two bayonets in the plurality of bayonets;

and determining a first-order adjacent bayonet shortest travel time matrix of the plurality of bayonets according to the first-order bayonet adjacent matrix and the shortest vehicle passing time between any two adjacent bayonets in the plurality of bayonets.

5. The method according to claim 1, wherein a first one of the mount identifiers of the snapshot sequence is preset as a current mount identifier, and a reasonable track list, an abnormal mount list and a suspended mount list are preset, wherein the reasonable track list is used for storing the target mount identifier group, and the suspended mount list is used for storing suspended mount identifiers;

the step of acquiring all target gate identifier groups in the snapshot sequence includes:

acquiring the current bayonet identification and a first bayonet identification adjacent to the current bayonet identification;

judging whether the difference value between the two capturing times respectively corresponding to the current bayonet identification and the first bayonet identification is greater than or equal to the shortest passing time between the two bayonets respectively corresponding to the current bayonet identification and the first bayonet identification;

if so, copying the current bayonet identification and the first bayonet identification to a target bayonet identification group to which the current bayonet identification belongs;

if not, executing a first hanging bayonet judgment step;

and taking the first gate identifier as the current gate identifier, and repeatedly executing the steps until the snapshot sequence is traversed, so as to obtain all target gate identifier groups in the snapshot sequence.

6. The method of claim 5, wherein the performing a first suspension bayonet determination step comprises:

judging whether a first target hanging bayonet mark exists in the hanging bayonet list or not; the difference value of the two capturing time corresponding to the first target hanging bayonet identification and the first bayonet identification is larger than or equal to the shortest passing time between the two bayonets corresponding to the first target hanging bayonet identification and the first bayonet identification respectively;

if so, copying the first bayonet identification to a target bayonet identification group to which the first target suspension bayonet identification belongs, and deleting the first target suspension bayonet; copying the current bayonet identification as a suspended bayonet identification to the suspended bayonet list; taking the first bayonet identification as the current bayonet identification, and returning to execute the step of acquiring the current bayonet identification and the first bayonet identification adjacent to the current bayonet identification;

and if not, executing an abnormal bayonet judgment step if the snapshot sequence is not traversed.

7. The method of claim 6, wherein the step of performing an abnormal bayonet determination comprises:

acquiring a second bayonet identification adjacent to the first bayonet identification;

judging whether the difference value between the two capturing times respectively corresponding to the current bayonet identification and the second bayonet identification is greater than or equal to the shortest passing time between the two bayonets respectively corresponding to the current bayonet identification and the second bayonet identification;

if yes, copying the second bayonet identification to a target bayonet identification group to which the current bayonet identification belongs, and copying the first bayonet identification to the abnormal bayonet list; taking the second bayonet identification as the current bayonet identification, and returning to execute the step of acquiring the current bayonet identification and the first bayonet identification adjacent to the current bayonet identification;

and if not, executing a second hanging bayonet judgment step.

8. The method of claim 7, wherein the performing a second suspension bayonet determination step comprises:

judging whether a second target hanging bayonet mark exists in the hanging bayonet list or not; the difference value of the two capturing times respectively corresponding to the second target hanging bayonet identification and the second bayonet identification is greater than or equal to the shortest passing time between the two bayonets respectively corresponding to the second target hanging bayonet identification and the second bayonet identification;

if so, copying the second bayonet identification to a target bayonet identification group to which the second target suspension bayonet belongs, and deleting the second target suspension bayonet; copying the first bayonet identification to the abnormal bayonet list; taking the second bayonet identification as the current bayonet identification, and returning to execute the step of acquiring the current bayonet identification and the first bayonet identification adjacent to the current bayonet identification;

if not, executing a second track identification step.

9. The method of claim 8, wherein the performing a second trajectory recognition step comprises:

judging whether the difference value of the two capturing times respectively corresponding to the first bayonet identification and the second bayonet identification is greater than or equal to the shortest passing time between the two bayonets respectively corresponding to the first bayonet identification and the second bayonet identification;

if so, adding a target bayonet identification group in the reasonable track list, and copying the first bayonet identification and the second bayonet identification to the added target bayonet identification group; taking the second bayonet identification as the current bayonet identification, and returning to execute the step of acquiring the current bayonet identification and the first bayonet identification adjacent to the current bayonet identification;

and if not, copying the first bayonet identification to the abnormal bayonet list, taking the second bayonet identification as the current bayonet identification, and returning to execute the step of acquiring the current bayonet identification and the first bayonet identification adjacent to the current bayonet identification.

10. A fake plate recognition device of a license plate is characterized by comprising:

11. A storage medium having a computer program stored thereon, the computer program, when being executed by a processor, realizing the steps of the method of any one of claims 1 to 9.

12. An electronic device comprising a processor and a memory, the memory storing a computer program, wherein the processor, when executing the computer program, implements the steps of the method of any one of claims 1 to 9.