CN105702047A - License plate identification error filtering method and apparatus in fake-license plate analysis - Google Patents
License plate identification error filtering method and apparatus in fake-license plate analysis Download PDFInfo
- Publication number
- CN105702047A CN105702047A CN201610124462.3A CN201610124462A CN105702047A CN 105702047 A CN105702047 A CN 105702047A CN 201610124462 A CN201610124462 A CN 201610124462A CN 105702047 A CN105702047 A CN 105702047A
- Authority
- CN
- China
- Prior art keywords
- car
- plate
- bayonet socket
- similar
- doubtful
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Devices For Checking Fares Or Tickets At Control Points (AREA)
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a license plate identification error filtering method and apparatus in fake-license plate analysis. The method includes the steps of: obtaining a suspected license plate number according to the fake-license plate analysis; obtaining all the similar license plates corresponding to the suspected fake-license plate; obtaining all inlet gates of a target gate through gate directed digraphs; if vehicle-passing records of the similar license plate of the suspected fake-license plate appear in some inlet gate of the target gate, the vehicle-passing records of the similar fake-license plate do not appear in all the outlet gates of the inlet gate, representing the suspected fake-license plate to be an identification error; filtering the suspected fake-license plate; and artificially checking the filtered suspected fake-license plate after the suspected fake-license plate is filtered, and finally confirming whether the license plate is fake. The apparatus of the invention includes a similar license plate analysis module, a first analysis filtering module, and a second analysis filtering module. The influence of license plate identification error is effectively reduced by performing identification error filtering on the suspected fake-license plate, the work load of the final artificial check is greatly reduced, and the final fake-license plate identification accuracy is improved.
Description
Technical field
The invention belongs to intelligent transportation field, particularly relate to Car license recognition mistake filter method and device during a kind of deck is analyzed。
Background technology
For improving traffic administration and public security regulatory level, it is ensured that the Effec-tive Function of road and the comprehensive management and control of traffic, the bayonet system based on video image identification is widely used。Each vehicular traffic is taken photo by the bayonet socket phase chance on road surface, and automatically by image intelligent analytical technology, obtain the information of vehicles such as the number-plate number after identifying, vehicle color, it is sent to central database to carry out unifying storage, based on these information of vehicles and other information, can completing the multiple business of traffic administration, public security management and control, namely the identification of fake-licensed car be one。
The recognition methods of existing fake-licensed car mainly has two kinds:
1, by analyzing the current record of the identical car plate of continuous print, see whether the connectedness of each bayonet socket point recording this car plate rationally determines whether deck。The method needs to be previously obtained the connected relation figure of all bayonet sockets in certain region, the comparison car plate to be analyzed current record at each bayonet socket is carried out with bayonet socket connected relation figure when analyzing deck, if car plate to be analyzed does not correspond with bayonet socket connected relation figure at the current record of each bayonet socket, then think there is deck, otherwise it is assumed that be reasonably cross car record。
2, by analyzing the current record of the identical car plate of continuous print, see vehicle that this identical car plate is corresponding is whether card inlet and the transit time interval gone out between bayonet socket rationally determine whether deck。The method needs to be previously obtained the connected relation of all bayonet sockets in certain region and from each card inlet to the shortest transit time going out bayonet socket, if vehicle corresponding to identical car plate at card inlet and the transit time that goes out between bayonet socket less than the shortest transit time between these two bayonet sockets, then think and occur in card inlet and the card release vehicle in car record of making a slip of the tongue is not same car, there is deck, otherwise it is assumed that be reasonably cross car record。
Both of these approaches is when practical application, owing to there is Car license recognition mistake, there will be the situation that two similar Car license recognition are become same car plate, if Car license recognition mistake is not filtered, to cause that final analysis fake license plate vehicle major part out is all the car plate identifying mistake, the accuracy rate of deck identification is very low, greatly increases the workload finally carrying out artificial nucleus couple。
Summary of the invention
It is an object of the invention to provide Car license recognition mistake filter method and device during a kind of deck is analyzed, the doubtful car plate identified during to traffic administration video analysis is identified mistake filtration, the problem not high to solve prior art traffic administration video analysis accuracy rate。
To achieve these goals, technical solution of the present invention is as follows:
Car license recognition mistake filter method in a kind of deck analysis, the doubtful number-plate number that deck analysis is obtained carries out Car license recognition mistake filtration, and described Car license recognition mistake filter method includes:
According to the doubtful number-plate number that deck analysis obtains, obtain all similar car plate that this doubtful number-plate number is corresponding；
What analyze the doubtful number-plate number in deck analysis goes out bayonet socket as target bayonet socket, all card inlets of this target bayonet socket are obtained from bayonet socket directed graph, within the time period set, the each similar car plate of the doubtful number-plate number is inquired about whether all card inlets exist the car record excessively of similar car plate, if what there is similar car plate crosses car record, then enter next step to continue to analyze, otherwise the doubtful number-plate number is not filtered；
For there is the card inlet crossing car record of similar car plate, obtain from bayonet socket directed graph exist similar car plate cross the card inlet of car record corresponding all go out bayonet socket, within the time period set, inquire about all go out bayonet socket whether exist this similar car plate cross car record, if what there is similar car plate crosses car record, then the doubtful number-plate number is not filtered, otherwise as identifying that mistake filters out this doubtful number-plate number。
Preferably, the described doubtful number-plate number obtained according to deck analysis, obtain all similar car plate that this doubtful number-plate number is corresponding, including:
Identification character according to the doubtful number-plate number and confidence level thereof, according to the confidence threshold value set, obtain the similar car plate that the doubtful number-plate number is corresponding。
Bayonet socket directed graph of the present invention is made up of summit and directed edge, summit is bayonet socket, and including card inlet and go out bayonet socket, directed edge represents the connected relation between two bayonet sockets, being pointed to out bayonet socket by card inlet, the weights of directed edge are the shortest transit time between two bayonet sockets。
Of the present invention within the time period set, the each similar car plate of the doubtful number-plate number is inquired about whether all card inlets exist the car record excessively of similar car plate, the time period wherein set refers to that initial time is as T α * tin, end time is the time period of T β * tin, wherein T is the time crossed in car record that on target bayonet socket, doubtful deck number is corresponding, tin is from card inlet to the shortest transit time of target bayonet socket, and α, β are parameter。
Of the present invention within the time period set, inquire about all go out bayonet socket whether exist this similar car plate cross car record, the time period wherein set refers to that initial time is as T '+β * tout, end time is the time period of T '+α * tout, wherein T ' is the time crossed in car record that on card inlet, similar car plate is corresponding, tout is the shortest transit time going out bayonet socket from this card inlet to correspondence, and α, β are parameter。
The invention allows for Car license recognition mistake defecator during a kind of deck is analyzed, the doubtful number-plate number that deck analysis is obtained carries out Car license recognition mistake filtration, it is characterised in that described Car license recognition mistake defecator includes:
Similar car plate analysis module, for the doubtful number-plate number obtained according to deck analysis, obtains all similar car plate that this doubtful number-plate number is corresponding；
First analyzes filtering module, bayonet socket is gone out as target bayonet socket for what analyze the doubtful number-plate number in deck analysis, all card inlets of this target bayonet socket are obtained from bayonet socket directed graph, within the time period set, the each similar car plate of the doubtful number-plate number is inquired about whether all card inlets exist the car record excessively of similar car plate, if what there is similar car plate crosses car record, then enter the second analysis filtering module and continue to analyze, otherwise the doubtful number-plate number is not filtered；
Second analyzes filtering module, for the card inlet crossing car record for there is similar car plate, obtain from bayonet socket directed graph exist similar car plate cross the card inlet of car record corresponding all go out bayonet socket, within the time period set, inquire about all go out bayonet socket whether exist this similar car plate cross car record, if what there is similar car plate crosses car record, then the doubtful number-plate number is not filtered, otherwise as identifying that mistake filters out this doubtful number-plate number。
Preferably, described similar car plate analysis module, when all similar car plates that the doubtful number-plate number of acquisition is corresponding, performs following operation:
Identification character according to the doubtful number-plate number and confidence level thereof, according to the confidence threshold value set, obtain the similar car plate that the doubtful number-plate number is corresponding。
The present invention proposes Car license recognition mistake filter method and device during a kind of deck is analyzed, in generating the process of similar car plate of doubtful deck, took full advantage of the recognition confidence of each character of car plate in car record, make the accuracy and efficiency of the similar car plate of generation be substantially improved；When carrying out doubtful deck and filtering in conjunction with bayonet socket directed graph and similar license plate number, employing time reasonability is filtered, and effectively reduces the impact of Car license recognition mistake, considerably reduces the quantity of doubtful deck, improve the efficiency of artificial nucleus couple, improve final deck recognition accuracy simultaneously。
Accompanying drawing explanation
A kind of deck that Fig. 1 provides for the embodiment of the present invention analyze in the flow chart of Car license recognition mistake filter method；
The schematic diagram of the bayonet socket directed graph that Fig. 2 provides for the embodiment of the present invention。
Detailed description of the invention
Below in conjunction with drawings and Examples, technical solution of the present invention being described in further details, following example do not constitute limitation of the invention。
Car license recognition mistake filter method in the analysis of the present embodiment one deck, is passing through the current record of the analysis identical car plate of continuous print, it is judged that after the car plate of doubtful deck, this car plate carries out Car license recognition mistake filtration。By obtaining the similar license plate number of doubtful deck, within the time period set, the car record excessively of similar license plate number is inquired about in conjunction with bayonet socket directed graph, by analyzing the car record of crossing of similar license plate number, doubtful fake license plate vehicle is carried out Car license recognition mistake filtration, promote the accuracy rate that deck is analyzed。
As it is shown in figure 1, Car license recognition mistake filter method in the analysis of the present embodiment one deck, including:
Step S1, the doubtful number-plate number obtained according to deck analysis, obtain all similar car plate that this doubtful number-plate number is corresponding。
In the present embodiment, bayonet socket camera is captured by the vehicle of bayonet socket, from the photo captured location, cut out car plate and be separately separated out by each character in car plate and be identified, identify the number-plate number obtaining vehicle。What form each vehicle crosses car record, crosses the number-plate number that car record includes the title of bayonet socket camera place bayonet socket, the photo of institute's take vehicle, bayonet socket camera the shooting time of photo, bayonet socket camera identify。
Deck analysis can be passed through to analyze the current record of the identical car plate of continuous print, see whether the connectedness of each bayonet socket point recording this car plate rationally determines whether deck, can also pass through to analyze the current record of the identical car plate of continuous print, see vehicle that this identical car plate is corresponding is whether card inlet and the transit time interval gone out between bayonet socket rationally determine whether deck。
Such as, as in figure 2 it is shown, pass through to configure or history is crossed the analysis of car record, obtain the rationally current shortest time of the oriented link relation between bayonet socket and correspondence, forming a bayonet socket directed graph, with bayonet socket for summit, the weights of directed edge are the shortest reasonable transit time between two bayonet sockets。
Such that it is able to based on whether bayonet socket Directed Graph analysis exists deck suspicion, for instance for all car records excessively to be analyzed, arrange according to shooting time ascending order。Analyze continuous two mistake car records of same car plate one by one, if the car time differences excessively of two records are more than or equal to the shortest transit time between the two bayonet socket, then it is assumed that be not deck, otherwise it is assumed that this car plate is doubtful deck。
In actual application, due to stained, light or other artificially block, there is the situation of wrong identification in the number-plate number identified, owing to car plate wrong identification easily causes the appearance of substantial amounts of doubtful deck situation。Filter it is thus desirable to carry out follow-up Car license recognition mistake according to the similar car plate that the number-plate number identified is corresponding。
When Car license recognition, the character to be identified separated in car plate and standard character are compared, character to be identified is the recognition confidence of this character to be identified with the ratio of same pixel in standard character, is assured that this character to be identified according to recognition confidence。
Generally using the highest character of recognition confidence as the end value of character to be identified, for instance a character is identified as B, and confidence level is 80%；Being identified as 8, confidence level is 40% simultaneously；Being also identified as D, confidence level is 20% simultaneously。Then this character final is identified as B。Assuming that this character comes the 4th in car plate, the value that other characters all determine that, then the number-plate number that this Car license recognition goes out is assumed to be: Zhejiang A1B567。
For the number-plate number identified, it is determined that its similar car plate, it is possible to obtain similar car plate according to the identification character of the doubtful number-plate number and confidence level thereof, it is also possible to obtain similar car plate according to similar character list。
Such as above example, the character coming the 4th in car plate is identified as B, and confidence level is 80%；Being identified as 8, confidence level is 40% simultaneously；Being also identified as D, confidence level is 20% simultaneously。Then its similar car plate is: Zhejiang A18567, or Zhejiang A1D567。
Again such as, in similar character list, B and 8, D are similar, it is possible to obtaining its similar car plate is: Zhejiang A18567, or Zhejiang A1D567。
The invention is not restricted to the method that concrete character recognition obtains with similar car plate, it is possible to determine similar car plate according to confidence level, it is also possible to obtain similar car plate according to similar character list。
Preferably, the present embodiment receives characters on license plate and the confidence level thereof of bayonet socket camera identification vehicle, according to the confidence threshold value set, obtains the similar car plate that the doubtful number-plate number is corresponding。Character for confidence level 100% directly adopts the similar characters on license plate identifying the characters on license plate that obtains as output；For confidence level lower than 100%, but higher than the character of the confidence threshold value (such as 40%) set, generate similar car plate as similar character；For not considering lower than the character of the confidence threshold value set。Such as above example, only export similar car plate Zhejiang A18567, and filter out Zhejiang A1D567, so can reduce computation complexity and promote error correction accuracy rate。
Step S2, in deck analysis, analyze the doubtful number-plate number go out bayonet socket as target bayonet socket, all card inlets of this target bayonet socket are obtained from bayonet socket directed graph, within the time period set, the each similar car plate of the doubtful number-plate number is inquired about whether all card inlets exist the car record excessively of similar car plate, if what there is similar car plate crosses car record, then enter next step to continue to analyze, otherwise the doubtful number-plate number is not filtered。
When carrying out deck and analyzing, if the car time difference excessively of continuous two mistake car records is more than or equal to the shortest transit time between the two bayonet socket, then it is assumed that be not deck, otherwise it is assumed that this car plate is doubtful deck。What video analysis went out doubtful deck crosses car record place goes out bayonet socket as target bayonet socket, for instance the bayonet socket E in Fig. 2。This car time that crosses crossing in car record record is T, all card inlet Tin of target bayonet socket are obtained, for instance bayonet socket A, the B in Fig. 2 from bayonet socket directed graph, and the rationally current shortest time tin from each card inlet to target bayonet socket, such as t3, t4, be designated as TollgateIns={Tini,tini, wherein i represents i-th card inlet。
Inquire about each similar car plate [T α * tin on each card inlet of target bayonet socket one by onei,T–β*tini] whether there is car record in the time period, if certain card inlet of target bayonet socket is at [T α * tini,T–β*tini] have similar car plate in the time period cross car record, then proceed next step and analyze, enter step S3；If all card inlets of target bayonet socket are at [T α * tini,T–β*tini] the time period interior car record excessively all without similar car plate, then it is assumed that it is absent from identifying wrong situation, this doubtful deck is not filtered。Wherein time period [T α * tini,T–β*tini] refer to that initial time is T α * tini, the end time is T β * tiniTime period。
Wherein, α, β are parameter, and in the present embodiment, α is 2, β is 1。Namely inquiring about at the shortest transit time to crossing car record in rational transit time, to reduce the workload inquired about, wherein α, β can be arranged flexibly according to practical situation, for instance arranging α be 3, β is 1。
Step S3, for exist similar car plate cross car record card inlet, obtain from bayonet socket directed graph exist similar car plate cross the card inlet of car record corresponding all go out bayonet socket, in predetermined time period, inquire about all go out bayonet socket whether exist this similar car plate cross car record, if what there is similar car plate crosses car record, then the doubtful number-plate number is not filtered, otherwise as identifying that mistake filters out this doubtful number-plate number。
Obtain from bayonet socket directed graph occur similar car plate cross the card inlet of car record corresponding all go out bayonet socket Tout, and from card inlet to the rationally current shortest time tout going out bayonet socket, be designated as TollgateOuts={Touti,touti, wherein i represents that i-th goes out bayonet socket, and target bayonet socket to be analyzed is also contained in ToutiIn, the similar car plate time crossed in car record on this card inlet is T ', toutiIt is from this card inlet to the shortest transit time going out bayonet socket i。What such as card inlet A was corresponding go out bayonet socket includes bayonet socket C, D, E, it is necessary to inquire about all go out bayonet socket C, D, E。
[T '+β * tout, T '+α * tout] in the time period, namely it is T '+β * tout in initial time, end time is in the time period of T '+α * tout, inquire about this card inlet all go out bayonet socket whether have this similar car plate cross car record, if all go out bayonet socket at [the T+ β * tout of its correspondence, T+ α * tout] the license plate number identification mistake of doubtful deck in target bayonet socket then thought in car record of crossing all without similar car plate in the time period, think to identify mistake, this doubtful number-plate number should be filtered out, otherwise retain this doubtful number-plate number。
It should be noted that in the present embodiment, α is 2, β is 1, above-mentioned parameter α, β can be adjusted by configuring according to the jam situation of road, can adjust the value that α is 3 or bigger when blocking up very much。
Doubtful deck completes to identify that mistake exports the final all doubtful set trade mark after filtering, by manually carrying out last check and correction and confirmation。
Citing below illustrates:
As in figure 2 it is shown, the subnet that this figure is a bayonet socket directed graph, A～G is bayonet socket, and t1～t6 is the shortest time rationally current between two bayonet sockets；Car plate CarNum to be analyzed ascending order temporally has two continuous print to cross car record at A bayonet socket and E bayonet socket, is T in the time crossed in car record of E bayonet socket；
Car plate CarNum is calculated from A bayonet socket to the transit time of E bayonet socket by the car recording gauge of crossing of A bayonet socket and E bayonet socket, find that this transit time is less than t3 (from A bayonet socket to E bayonet socket short across the car time), then unreasonable on express time, this car plate CarNum is doubtful deck；
Carry out first obtaining when doubtful deck filters the similar license plate number CarNum ' of this doubtful set trade mark CarNum；
All card inlets of E bayonet socket are obtained, respectively A bayonet socket and B bayonet socket by bayonet socket Directed Graph analysis；From A bayonet socket to E bayonet socket is t3 short across the car time, and from B bayonet socket to E bayonet socket is t4 short across the car time；Inquiry [T 2*t3, T t3] in the time period car plate CarNum ' whether had a car record at A bayonet socket, inquiry [T 2*t4, T t4] in the time period car plate CarNum ' whether B bayonet socket had car record；By inquire about find [T 2*t3, T t3] in the time period A bayonet socket have car plate CarNum's ' to cross car record, the time crossed in car record is T '；
By bayonet socket Directed Graph analysis obtain A bayonet socket all go out bayonet socket, respectively C bayonet socket, D bayonet socket and E bayonet socket；A bayonet socket is t1 short across the car time to C bayonet socket, and A bayonet socket is t2 short across the car time to D bayonet socket, and A bayonet socket is t3 short across the car time to E bayonet socket；Find by inquiring about, [T '+t1, T '+2*t1] in the time period, C bayonet socket does not have CarNum's ' to cross car record, [T '+t2, T '+2*t2] in the time period, D bayonet socket does not have CarNum's ' to cross car record, [T '+t3, T '+2*t3] time period E bayonet socket do not have CarNum ' cross car record, then think and occur in that Car license recognition mistake, actual the T moment occur in E bayonet socket be not car plate CarNum but car plate CarNum ', doubtful deck CarNum is filtered out。
Car license recognition mistake defecator in the analysis of the present embodiment one deck, corresponding with said method, the doubtful number-plate number that deck analysis is obtained carries out Car license recognition mistake filtration, and this Car license recognition mistake defecator includes:
Similar car plate analysis module, for the doubtful number-plate number obtained according to deck analysis, obtains all similar car plate that this doubtful number-plate number is corresponding；
First analyzes filtering module, bayonet socket is gone out as target bayonet socket for what analyze the doubtful number-plate number in deck analysis, all card inlets of this target bayonet socket are obtained from bayonet socket directed graph, within the time period set, the each similar car plate of the doubtful number-plate number is inquired about whether all card inlets exist the car record excessively of similar car plate, if what there is similar car plate crosses car record, then enter the second analysis filtering module and continue to analyze, otherwise the doubtful number-plate number is not filtered；
Second analyzes filtering module, for the card inlet crossing car record for there is similar car plate, obtain from bayonet socket directed graph exist similar car plate cross the card inlet of car record corresponding all go out bayonet socket, within the time period set, inquire about all go out bayonet socket whether exist this similar car plate cross car record, if what there is similar car plate crosses car record, then the doubtful number-plate number is not filtered, otherwise as identifying that mistake filters out this doubtful number-plate number。
Preferably, similar car plate analysis module, when all similar car plates that the doubtful number-plate number of acquisition is corresponding, performs following operation:
Identification character according to the doubtful number-plate number and confidence level thereof, according to the confidence threshold value set, obtain the similar car plate that the doubtful number-plate number is corresponding。
Above example is only limited in order to technical scheme to be described; when without departing substantially from present invention spirit and essence thereof; those of ordinary skill in the art are when can make various corresponding change and deformation according to the present invention, but these change accordingly and deformation all should belong to the scope of the claims appended by the present invention。
Claims (10)
1. a Car license recognition mistake filter method in deck analysis, the doubtful number-plate number that deck analysis is obtained carries out Car license recognition mistake filtration, it is characterised in that described Car license recognition mistake filter method includes:
According to the doubtful number-plate number that deck analysis obtains, obtain all similar car plate that this doubtful number-plate number is corresponding；
What analyze the doubtful number-plate number in deck analysis goes out bayonet socket as target bayonet socket, all card inlets of this target bayonet socket are obtained from bayonet socket directed graph, within the time period set, the each similar car plate of the doubtful number-plate number is inquired about whether all card inlets exist the car record excessively of similar car plate, if what there is similar car plate crosses car record, then enter next step to continue to analyze, otherwise the doubtful number-plate number is not filtered；
For there is the card inlet crossing car record of similar car plate, obtain from bayonet socket directed graph exist similar car plate cross the card inlet of car record corresponding all go out bayonet socket, within the time period set, inquire about all go out bayonet socket whether exist this similar car plate cross car record, if what there is similar car plate crosses car record, then the doubtful number-plate number is not filtered, otherwise as identifying that mistake filters out this doubtful number-plate number。
2. Car license recognition mistake filter method according to claim 1, it is characterised in that the described doubtful number-plate number obtained according to deck analysis, obtains all similar car plate that this doubtful number-plate number is corresponding, including:
Identification character according to the doubtful number-plate number and confidence level thereof, according to the confidence threshold value set, obtain the similar car plate that the doubtful number-plate number is corresponding。
3. Car license recognition mistake filter method according to claim 1, it is characterized in that, described bayonet socket directed graph is made up of summit and directed edge, summit is bayonet socket, including card inlet and go out bayonet socket, directed edge represents the connected relation between two bayonet sockets, card inlet point to out bayonet socket, and the weights of directed edge are the shortest transit time between two bayonet sockets。
4. Car license recognition mistake filter method according to claim 1, it is characterized in that, described within the time period set, the each similar car plate of the doubtful number-plate number is inquired about whether all card inlets exist the car record excessively of similar car plate, the time period wherein set refers to that initial time is as T α * tin, end time is the time period of T β * tin, wherein T is the time crossed in car record that on target bayonet socket, doubtful deck number is corresponding, tin is from card inlet to the shortest transit time of target bayonet socket, and α, β are parameter。
5. Car license recognition mistake filter method according to claim 1, it is characterized in that, described within the time period set, inquire about all go out bayonet socket whether exist this similar car plate cross car record, the time period wherein set refers to that initial time is as T '+β * tout, and the end time is the time period of T '+α * tout, and wherein T ' is the time crossed in car record that on card inlet, similar car plate is corresponding, tout is the shortest transit time going out bayonet socket from this card inlet to correspondence, and α, β are parameter。
6. a Car license recognition mistake defecator in deck analysis, the doubtful number-plate number that deck analysis is obtained carries out Car license recognition mistake filtration, it is characterised in that described Car license recognition mistake defecator includes:
Similar car plate analysis module, for the doubtful number-plate number obtained according to deck analysis, obtains all similar car plate that this doubtful number-plate number is corresponding；
First analyzes filtering module, bayonet socket is gone out as target bayonet socket for what analyze the doubtful number-plate number in deck analysis, all card inlets of this target bayonet socket are obtained from bayonet socket directed graph, within the time period set, the each similar car plate of the doubtful number-plate number is inquired about whether all card inlets exist the car record excessively of similar car plate, if what there is similar car plate crosses car record, then enter the second analysis filtering module and continue to analyze, otherwise the doubtful number-plate number is not filtered；
Second analyzes filtering module, for the card inlet crossing car record for there is similar car plate, obtain from bayonet socket directed graph exist similar car plate cross the card inlet of car record corresponding all go out bayonet socket, within the time period set, inquire about all go out bayonet socket whether exist this similar car plate cross car record, if what there is similar car plate crosses car record, then the doubtful number-plate number is not filtered, otherwise as identifying that mistake filters out this doubtful number-plate number。
7. Car license recognition mistake defecator according to claim 6, it is characterised in that described similar car plate analysis module, when all similar car plates that the doubtful number-plate number of acquisition is corresponding, performs following operation:
Identification character according to the doubtful number-plate number and confidence level thereof, according to the confidence threshold value set, obtain the similar car plate that the doubtful number-plate number is corresponding。
8. Car license recognition mistake defecator according to claim 6, it is characterized in that, described bayonet socket directed graph is made up of summit and directed edge, summit is bayonet socket, including card inlet and go out bayonet socket, directed edge represents the connected relation between two bayonet sockets, card inlet point to out bayonet socket, and the weights of directed edge are the shortest transit time between two bayonet sockets。
9. Car license recognition mistake defecator according to claim 6, it is characterized in that, described within the time period set, the each similar car plate of the doubtful number-plate number is inquired about whether all card inlets exist the car record excessively of similar car plate, the time period wherein set refers to that initial time is as T α * tin, end time is the time period of T β * tin, wherein T is the time crossed in car record that on target bayonet socket, doubtful deck number is corresponding, tin is from card inlet to the shortest transit time of target bayonet socket, and α, β are parameter。
10. Car license recognition mistake defecator according to claim 6, it is characterized in that, described within the time period set, inquire about all go out bayonet socket whether exist this similar car plate cross car record, the time period wherein set refers to that initial time is as T '+β * tout, and the end time is the time period of T '+α * tout, and wherein T ' is the time crossed in car record that on card inlet, similar car plate is corresponding, tout is the shortest transit time going out bayonet socket from this card inlet to correspondence, and α, β are parameter。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610124462.3A CN105702047B (en) | 2016-03-04 | 2016-03-04 | Car license recognition mistake filter method and device in a kind of deck analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610124462.3A CN105702047B (en) | 2016-03-04 | 2016-03-04 | Car license recognition mistake filter method and device in a kind of deck analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105702047A true CN105702047A (en) | 2016-06-22 |
| CN105702047B CN105702047B (en) | 2018-06-05 |
Family
ID=56220727
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610124462.3A Active CN105702047B (en) | 2016-03-04 | 2016-03-04 | Car license recognition mistake filter method and device in a kind of deck analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105702047B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106571040A (en) * | 2016-10-31 | 2017-04-19 | 浙江宇视科技有限公司 | Suspicious vehicle confirmation method and equipment |
| CN107329977A (en) * | 2017-05-27 | 2017-11-07 | 银江股份有限公司 | A kind of false-trademark car postsearch screening method based on probability distribution |
| CN107730625A (en) * | 2017-06-08 | 2018-02-23 | 西安艾润物联网技术服务有限责任公司 | Intelligent parking system goes out car charging method, system and storage medium |
| CN107862072A (en) * | 2017-11-24 | 2018-03-30 | 泰华智慧产业集团股份有限公司 | Enter the method for city deck crime based on big data technical Analysis vehicle |
| CN108073923A (en) * | 2016-11-09 | 2018-05-25 | 杭州海康威视数字技术股份有限公司 | A kind of VLP correction method and device |
| CN108197625A (en) * | 2017-12-18 | 2018-06-22 | 北京云星宇交通科技股份有限公司 | A kind of method and system for correcting Car license recognition |
| CN109272034A (en) * | 2018-09-07 | 2019-01-25 | 佳都新太科技股份有限公司 | A kind of missing inspection license plate prediction technique and processing terminal |
| CN110148300A (en) * | 2019-06-24 | 2019-08-20 | 上海擎感智能科技有限公司 | A kind of counterfeit vehicle registration plate identification method, device and computer-readable medium |
| CN111223305A (en) * | 2019-12-25 | 2020-06-02 | 华为技术有限公司 | Method and device for detecting illegal card involvement |
| CN112348010A (en) * | 2019-08-07 | 2021-02-09 | 杭州海康威视系统技术有限公司 | License plate matching method and device |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102945605A (en) * | 2012-11-06 | 2013-02-27 | 东软集团股份有限公司 | Fake-licensed car identification method and system |
| CN103116989A (en) * | 2013-01-29 | 2013-05-22 | 江苏省邮电规划设计院有限责任公司 | Smart city based fake platevehicle capture method |
| CN103413440A (en) * | 2013-04-11 | 2013-11-27 | 江苏省邮电规划设计院有限责任公司 | Fake-licensed vehicle identification method based on smart city data base and identification rule base |
| CN103700262A (en) * | 2013-12-26 | 2014-04-02 | 福建省视通光电网络有限公司 | Automatic area acquisition method for fake-licensed vehicles |
| CN103914986A (en) * | 2014-03-14 | 2014-07-09 | 浙江宇视科技有限公司 | Method and device for fake-license-plate analysis |
| CN104021677A (en) * | 2014-06-25 | 2014-09-03 | 浙江宇视科技有限公司 | Method and device for detecting turnover of number plate |
| CN104517445A (en) * | 2014-12-09 | 2015-04-15 | 浪潮电子信息产业股份有限公司 | Fake-licensed vehicle identification method based on big data |
-
2016
- 2016-03-04 CN CN201610124462.3A patent/CN105702047B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102945605A (en) * | 2012-11-06 | 2013-02-27 | 东软集团股份有限公司 | Fake-licensed car identification method and system |
| CN103116989A (en) * | 2013-01-29 | 2013-05-22 | 江苏省邮电规划设计院有限责任公司 | Smart city based fake platevehicle capture method |
| CN103413440A (en) * | 2013-04-11 | 2013-11-27 | 江苏省邮电规划设计院有限责任公司 | Fake-licensed vehicle identification method based on smart city data base and identification rule base |
| CN103700262A (en) * | 2013-12-26 | 2014-04-02 | 福建省视通光电网络有限公司 | Automatic area acquisition method for fake-licensed vehicles |
| CN103914986A (en) * | 2014-03-14 | 2014-07-09 | 浙江宇视科技有限公司 | Method and device for fake-license-plate analysis |
| CN104021677A (en) * | 2014-06-25 | 2014-09-03 | 浙江宇视科技有限公司 | Method and device for detecting turnover of number plate |
| CN104517445A (en) * | 2014-12-09 | 2015-04-15 | 浪潮电子信息产业股份有限公司 | Fake-licensed vehicle identification method based on big data |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106571040A (en) * | 2016-10-31 | 2017-04-19 | 浙江宇视科技有限公司 | Suspicious vehicle confirmation method and equipment |
| CN108073923B (en) * | 2016-11-09 | 2021-09-17 | 杭州海康威视数字技术股份有限公司 | License plate correction method and device |
| CN108073923A (en) * | 2016-11-09 | 2018-05-25 | 杭州海康威视数字技术股份有限公司 | A kind of VLP correction method and device |
| US11023759B2 (en) | 2016-11-09 | 2021-06-01 | Hangzhou Hikvision Digital Technology Co., Ltd. | Method and device for correcting license plate |
| CN107329977B (en) * | 2017-05-27 | 2019-08-16 | 银江股份有限公司 | A kind of false-trademark vehicle postsearch screening method based on probability distribution |
| CN107329977A (en) * | 2017-05-27 | 2017-11-07 | 银江股份有限公司 | A kind of false-trademark car postsearch screening method based on probability distribution |
| CN107730625A (en) * | 2017-06-08 | 2018-02-23 | 西安艾润物联网技术服务有限责任公司 | Intelligent parking system goes out car charging method, system and storage medium |
| CN107862072B (en) * | 2017-11-24 | 2020-05-19 | 泰华智慧产业集团股份有限公司 | Method for analyzing vehicle urban-entering fake plate crime based on big data technology |
| CN107862072A (en) * | 2017-11-24 | 2018-03-30 | 泰华智慧产业集团股份有限公司 | Enter the method for city deck crime based on big data technical Analysis vehicle |
| CN108197625A (en) * | 2017-12-18 | 2018-06-22 | 北京云星宇交通科技股份有限公司 | A kind of method and system for correcting Car license recognition |
| CN109272034A (en) * | 2018-09-07 | 2019-01-25 | 佳都新太科技股份有限公司 | A kind of missing inspection license plate prediction technique and processing terminal |
| CN110148300A (en) * | 2019-06-24 | 2019-08-20 | 上海擎感智能科技有限公司 | A kind of counterfeit vehicle registration plate identification method, device and computer-readable medium |
| CN112348010A (en) * | 2019-08-07 | 2021-02-09 | 杭州海康威视系统技术有限公司 | License plate matching method and device |
| CN111223305A (en) * | 2019-12-25 | 2020-06-02 | 华为技术有限公司 | Method and device for detecting illegal card involvement |
| CN111223305B (en) * | 2019-12-25 | 2022-05-17 | 华为技术有限公司 | Method and device for detecting illegal card involvement |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105702047B (en) | 2018-06-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105702047A (en) | License plate identification error filtering method and apparatus in fake-license plate analysis | |
| CN110490108B (en) | Violation state marking method and device, storage medium and electronic device | |
| KR101988540B1 (en) | System and method for managing parking | |
| CN105488479B (en) | A kind of counterfeit vehicle registration plate identification method and device based on weighting local feature comparison | |
| CN103914986B (en) | The method and apparatus that a kind of deck is analyzed | |
| CN111161543A (en) | Automatic snapshot method and system for bus front violation behavior based on image recognition | |
| CN105488453A (en) | Detection identification method of no-seat-belt-fastening behavior of driver based on image processing | |
| CN103246876A (en) | Image feature comparison based counterfeit vehicle registration plate identification method | |
| CN105335710A (en) | Fine vehicle model identification method based on multi-stage classifier | |
| CN111815959A (en) | Vehicle violation detection method and device and computer readable storage medium | |
| CN106934378B (en) | Automobile high beam identification system and method based on video deep learning | |
| WO2017125063A1 (en) | Processing method and device for vehicle traffic violation | |
| CN105160884A (en) | Vehicle identification method and device fusing video monitoring and RFID | |
| CN104794931A (en) | Method and system for monitoring usage of parking space through cameras | |
| CN108765975A (en) | The vertical managing system of car parking of trackside and method | |
| CN106960579A (en) | The recognition methods of unlicensed car a kind of and use its parking system | |
| CN109523792A (en) | Automatic correcting method, medium and the system of license plate recognition result | |
| CN110310378A (en) | A kind of open type parking ground parking charge method and system based on double mirror | |
| DE102017209496A1 (en) | Method and device for classifying an object for a vehicle | |
| CN107331206A (en) | The Car license recognition intelligent parking management system and method aided in based on bluetooth | |
| CN103810298A (en) | Main memory database based rapid vehicle information query method and system | |
| CN112115939A (en) | Vehicle license plate recognition method and device | |
| Agbeyangi et al. | Automatic identification of vehicle plate number using raspberry pi | |
| CN103106400B (en) | A kind of method for detecting human face and device | |
| CN118230563A (en) | Parking space lock management method and related device based on license plate recognition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |