CN101097605A - Vehicle personal identification system of ETC electric no-parking charge - Google Patents
Vehicle personal identification system of ETC electric no-parking charge Download PDFInfo
- Publication number
- CN101097605A CN101097605A CNA2006100180461A CN200610018046A CN101097605A CN 101097605 A CN101097605 A CN 101097605A CN A2006100180461 A CNA2006100180461 A CN A2006100180461A CN 200610018046 A CN200610018046 A CN 200610018046A CN 101097605 A CN101097605 A CN 101097605A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- information
- identification
- payment
- chess
- 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
Images
Abstract
The invention discloses a kind of recognition system of vehicle ID of ETC electron without stopping to charging, which includes IC card of vehicle-carrying radio frequency containing the basic information of vehicle and several information channels connect with several sensors, and the information channels is combined by anabranch model of play chess information and the ratiocination algorithm of Bayes's network to complete the precious identification for vehicle ID. The invention anabranches the information of vehicle ID in non-contact intelligent IC card based on the anabranch theory of playing chess information, and the several source information can reach the optimized reliable judgment by the competition and the corporation in the anabranch model, and identify the real ID of vehicle, and ensures the reliability of identification, and the precision can reach above 99.99%, and the invest is little, and the reliability is well. The recognition system of vehicle ID does not occupy the source of host computer and improve the work efficiency of host computer. The identifying device is small and easy to install.
Description
Technical field
The present invention relates to the vehicle personal identification system in a kind of ETC electric non-stop toll.
Background technology
ETC electric non-stop toll technology is since the initial stage nineties occurs, successively in the North America, ground widespread uses such as Europe, Japan, Australia, Singapore, China has organized " the no-stop charging system industry is united tackling key problem under the network environment " project research also during the State's Ninth Five-Year Plan period by Ministry of Communications.Bear the scientific research mission in early stage by Research Institute of Highway, Ministry of Communications, Xi'an Science Institute, and launch the demonstration project construction in Department of Communications of Guangdong Province, Beijing transportation department, Jiangsu Province transportation department etc.Meanwhile, each big institution of higher learning of the whole nation, research institution have also carried out ETC electric non-stop toll Study on Technology one after another.
But because the China's national situation difference, the reason of the imperfection of National Credit system and China's expressway tol lcollection system makes ETC electric non-stop toll technology run into very big difficulty applying of China.The ETC electronic charging system without parking that various places (as Guangdong, Chongqing, Shanghai and other places) have started is stopped using one after another or is changed and does its usefulness.Although it is many-sided causing the reason of this situation, but blocking, change card, charge vehicle uses free card, oversize vehicle to use a large amount of existence of bad and swindle phenomenon such as dilly rate card, being the subject matter that the ETC electric non-stop toll faces, is the main cause that causes the ETC electric non-stop toll can not effectively promote in China.For example, on national expressway network toll technical seminar in 2004, point out with regard to the representative that Chongqing is arranged that Chongqing is at the initial stage of implementing the ETC electric non-stop toll, in one month, the generation of above-mentioned various phenomenons is just up to up to a hundred.
Blocking, change card, charge vehicle and use free card, oversize vehicle to use a large amount of existence of bad and swindle phenomenon such as dilly rate card, all is because existing ETC electronic charging system without parking can not accurately be discerned the identity of vehicle effectively causes in essence.Because traditional ETC electronic charging system without parking, with the foundation of IC-card as the testing vehicle register differentiation, but IC-card is movable, do not fix onboard, whether a vehicle-mounted IC-card is the IC-card of this car really just, in the ETC system, other auxiliary evidence does not confirm, has therefore caused a large amount of existence of above-mentioned phenomenon.We can say that the accurate decision problem of testing vehicle register has become China ETC electric non-stop toll so that the bottleneck of intelligent transportation transportation development.
In fact, some domestic research institutions have realized that this problem and study, and have proposed some schemes and implement.For example, by license plate identification, compare with the licence plate that writes down in the vehicle-mounted IC card; Or utilize magnetic induction coil, piezoelectric sensor, far infrared transceiver etc. to judge that vehicle arrives, and obtain wheel number, front and rear wheel distance, the wheelbase of traffick, features such as overall height realize vehicle differentiation etc. then.But these schemes all can not be effectively, deal with problems up hill and dale.Analyze its reason, mainly contain following two aspects:
(1) precision of car plate identification does not reach the requirement of application.Though the research of the domestic existing a lot of units of car plate identification, technical also comparative maturity, still very undesirable from the situation of practical application.Because on-the-spot car plate environment-identification is very complicated, be subjected to influence of various factors such as car plate s own situation, weather, illumination, some car plate even people discern difficulties.For example, the Vehicle License Plate Recognition System (Han Wang Eye numberplate recognizing system) of domestic state-of-the-art Beijing Chinese princes and dukes department, when parameters is adjusted to optimum condition, on-the-spot car plate discrimination also has only 96%, but the adjustment of parameter will be adjusted several times according to weather, illumination etc. in one day repeatedly, otherwise the car plate discrimination will reduce greatly, cause the practicality of this system to be had a greatly reduced quality.Chongqing is when implementing non-parking charge, because car plate identification can not reach application requirements, have to behind the artificial input of outlet car plate four, with the car plate data comparison of writing down in the IC-card, but so, vehicle just must stop more than 5 seconds in outlet, and non-parking charge returns to Car park payment again and got on.In addition, the ETC non-parking charge in Shanghai also is because the car plate accuracy of identification is not high and stranded.Because car plate identification can not reach application requirements, cause in the checking of testing vehicle register, can not judge is the illegal IC-card or the reason of car plate identification on earth, so only use the method validation testing vehicle register of car plate identification, is not effective, a reliable solution.
(2) the vehicle recognition result is too simple, is not enough to form the effective reliable authentication to testing vehicle register.Present vehicle identification is divided into hardware mode and software mode.Utilizing hardware mode to obtain vehicle needs equipment such as ground induction coil, magnetic induction coil, piezoelectric sensor, infrared facility, and the systematic comparison complexity is invested also greatlyyer, no longer adopts basically at present.Utilize software mode to be meant by taking vehicle image, image is carried out vehicle to be discerned and judges vehicle, this method only needs a camera to get final product, but recognition result generally has only three kinds of compact cars, in-between car, large car, and a lot of vehicles can't be from accurately distinguishing in appearance, as the differentiation of large and medium bus and motor van, cause with the vehicle recognition result as judging the poorer of testing vehicle register according to effect, particularly run into non-free car and use the situation of free card all the more so.
Because above two aspects, the problem that the testing vehicle register that makes restriction ETC promote in China is discerned can not effectively be solved all the time.In ETC, if testing vehicle register can not obtain effective recognition, the disastrous effect that causes is many-sided: at first can cause the inconsistent of testing vehicle register differentiation, derive from the result of data in the IC-card and licence plate or vehicle identification when inconsistent, both may be that illegal IC-card causes, the mistake identification that also may be car plate or vehicle causes, this situation often needs to stop to handle, when owing to being the reason of car plate or vehicle identification, cause parking bring very big negative effect can for the reputation and the charge station of ETC electronic non-parking system; Secondly, testing vehicle register is differentiated mistake and both can be caused damage to the driver, or connives driver's profiteering attitude, also can cause ticket fee to run off to charge station.
Summary of the invention
The purpose of this invention is to provide the vehicle personal identification system in a kind of ETC electric non-stop toll that can accurately discern testing vehicle register.
For achieving the above object, the present invention adopts following technical scheme: the vehicle personal identification system in the ETC electric non-stop toll, comprise the vehicle-mounted radio-frequency (RF) IC card and the read-write system thereof that there are vehicle essential information data, contain a plurality of information channels that are connected with a plurality of sensors, each information channel is played chess information fusion model with fighting to be organized, with the reasoning algorithm of Bayesian network, finish accurate identification to testing vehicle register; Described fighting played chess information fusion model and comprised player, strategy and pay three fundamentals, the player is meant to fight and plays chess the participator of fusion, a plurality of information channels that promptly being used to fights plays chess fusion, strategy be to the player when the various situation that may occur in playing chess is fought in processing, a kind of description of the decision-making of doing, be the mode of operation or the running parameter of each information channel, payment is the degree of confidence of each information channel to the decision support of information fusion center.
Described fight play chess that information fusion model comprises that the player extracts, strategy extracts, payment is extracted, situation assessment, process extract five modules and data base management system (DBMS), described data base management system (DBMS) is made up of policy database and payment data storehouse, wherein policy database is used for the different players' of storage system admissible strategy, and the payoff of Different Strategies combination is then deposited in the payment data storehouse; Described reasoning algorithm is non-tree type Bayesian network reasoning algorithm.
The described a plurality of information channels that are used to merge comprise license board information and vehicle information, and described sensor is video camera and weighing system.
Described vehicle information comprises car plate color, vehicle and car weight, and wherein the car plate color comprises yellow, blue look, black, white, and it is large-scale, medium-sized, small-sized that vehicle comprises, that car weight comprises is very heavy, heavy, medium, gently, very light five quantized levels; License board information and car plate color, vehicle are all gathered as sensor by video camera, and car weight is gathered as sensor by weighing system.
License board information matching degree in described license plate image identification channel information and radio-frequency (RF) identification passage is big more, and this passage payment is big more; Vehicle information conforms degree in vehicle image recognition channel information and the radio-frequency (RF) identification passage is big more, and this passage payment is big more, but under identical payment, big not as good as license plate image identification channel information to the influence of emerging system decision-making; Vehicle weight information and car plate colouring information are all relevant with vehicle information in the radio-frequency (RF) identification passage, determine their payment size according to the size of correlativity.
The ETC electronic charging system without parking of playing chess information fusion technology based on fighting that we invented is intended to set up the information fusion theory problem system that solves under the multi-source information conflict environment, thereby has theory significance; On the other hand, under the guidance of this theoretical system, the Fusion Model of setting up is applied in the reliable decision of testing vehicle register in the ETC electric non-stop toll,, solves the reliable charging problem in the ETC electronic charging system without parking by setting up specific Fusion Model based on this theory.It plays chess information fusion model by fighting, the information fusion that the information of the information of the information in the contactless IC card, car plate identification, vehicle identification, car plate colouring information and parking vehicle are not weighed is in the same place, play chess competition and cooperation in the information fusion model by fighting in the hope of above-mentioned multi-source information and reach system optimal and reliably judge, confirm jointly or the true identity of the real vehicle of giving counterevidence.Because people know from general knowledge, if the evidence that a plurality of information channels provide combines, to have stronger proof than the evidence that single information channel provides, discern with regard to the testing vehicle register among the ETC, the decision-making that provides after the multiple information fusion---be legal vehicle or illegal vehicle, will be more more reliable than the judgement of only making by single information.Our target is: play chess the means of information fusion by fighting, solve testing vehicle register and differentiate unsolved always emphasis and difficult point problem in this ETC electric non-stop toll.The solution of key issue among this ETC, not only indicated direction in the trend of China's development for the ETC technology, and established theoretical foundation for the practicability of ETC system, this was a practical problems that derives from engineering background originally, but, it has been risen to fight and played chess the height of theoretical and Theory of Information Fusion by our research.
The beginning of this century, China has begun development and the pilot work of ETC, in August, 2004, Ministry of Communications's official approval " networking electronic charging working group " (UT﹠amp; ETC WG), be responsible for the standardization effort in networking electronic charging field, and the developing direction of electronic charging system without parking as the electronic charging of networking future, subsequently, the whole nation each relevant universities and colleges, company, research institution have begun the research of ETC one after another, and what have has also applied for patent.The ETC electronic charging system without parking of our development, in conjunction with China's national situation, employing is fought and is played chess information fusion technology identification vehicle true identity, the charge reliable basis is provided, consult pertinent literature through us, through the current awareness search work of Henan Province's Office for Scientific and Technical Information, similar technique and relevant report are not found in this technological invention.Specifically, our this technological invention has following advantage
1, it has satisfactorily resolved the problem of testing vehicle register need reliable recognition in the distinctive ETC of China.It is fought the weight information in car plate identification, vehicle identification, car plate color, IC-card data message and the weighing system that do not stop and plays chess information fusion, guaranteed the reliability of testing vehicle register identification to greatest extent, compared with traditional simple use license plate identification, precision is brought up to more than 99.99%, judge compared with simple use vehicle, small investment, good reliability, precision height.
2, adopt the mode of distributed treatment in the testing vehicle register judgement system, but car plate and vehicle identification and the independent operational system of employing of weighing of not stopping, do not take host resource, main frame can be concentrated carry out information fusion and charge data processing, improved the work efficiency of main frame so greatly, improved the pass through speed of vehicle in charge station.
3, owing to be that multiline message merges, the multichannel of information provides more redundant information, make when the loss of learning of a certain road, the information of other road operate as normal, still can improved blending algorithm, obtain higher differentiation accuracy rate, this just provides fault redundance for system, system still can be worked under the situation that local fault has taken place, thereby improved the reliability of system greatly, this point has just adapted to the requirement of ETC electronic charging system without parking needs higher reliabilities, becomes an obvious advantage of this technological invention.
4, it is little to be used to the equipment volume discerned, easy for installation, the employing portable system of weighing does not stop, incorporate system is used in car plate and vehicle identification, but therefore independent operating, weight information and recognition result insert simple through serial port and main-machine communication, easy for installation, saved the investment of user aspect engineering construction greatly.
Description of drawings
Figure 1 shows that the principle of work block diagram that comprises ETC of the present invention system;
Figure 2 shows that Vehicle License Plate Recognition System process flow diagram among the present invention;
Figure 3 shows that model recognition system process flow diagram among the present invention;
Figure 4 shows that fights among the present invention plays chess the functional mode of information fusion;
Figure 5 shows that the bayesian network structure figure of testing vehicle register identification among the present invention;
Figure 6 shows that the arthmetic statement figure of Bayesian network among the present invention;
Figure 7 shows that the algorithm routine process flow diagram of Bayesian network among the present invention;
Figure 8 shows that the radio-frequency recognition system program flow diagram that comprises in the ETC of the present invention system.
Embodiment
Vehicle personal identification system among the ETC is played chess information fusion to fight be guidance.In design, at first extract player, strategy, three of payments and fight and play chess the fundamental of Fusion Model, be to make up Bayesian network then, finish the identification of testing vehicle register with the reasoning algorithm of network.
(1) player
Directly with each information channel as the player.
(2) strategy
The different kinds of information passage is very big to the extraction influence of system strategy.Each player's is tactful as follows:
● the weighing system information channel of not stopping: the real number value of output.Need to adopt fuzzy technology carry out limited quantification, be divided into several quantized levels it: as very heavy, heavy, medium, gently, very light.
● license plate image information channel: the output result (character format) of camera parameters (focal length, the visual field, angle etc.), recognition system.
● vehicle image recognition passage: camera parameters is the same, identification output result's (large-scale, medium-sized, small-sized).
● car plate color information channel: yellow, blue look, black, white.
● radio-frequency (RF) identification passage: be stored in car plate, model data in the radio-frequency card.
(3) payment
This payment of playing chess each channel information of information fusion system of fighting is embodied on the matching degree to the radio-frequency (RF) identification channel information.Matching degree is big more, shows that the payment of information is big more.Make a concrete analysis of below:
● information in radio frequency is owing to be exactly itself, so matching degree is 100%, so in this emerging system not as decision-making foundation.
● the license board information matching degree in license plate image identification channel information and the radio-frequency (RF) identification passage is big more, and this passage payment is big more.
● the vehicle information conforms degree in vehicle image recognition channel information and the radio-frequency (RF) identification passage is big more, and this passage payment is big more, but under identical payment, big not as good as license plate image identification channel information to the influence of emerging system decision-making.
● vehicle weight information and car plate colouring information are all relevant with vehicle information in the radio-frequency (RF) identification passage, determine their payment size according to the size of correlativity.
Fight and play chess information fusion system and will make decisions: promptly be legal vehicle or illegal vehicle according to the payment of each information channel.The foundation that the payment of any single passage all can not be made a strategic decision as emerging system fully.In order to help the system decision-making, each passage all should be striven for the payment maximization, and emerging system is striven for the maximum utilization to each information channel, so each information channel exists the competition of payment and fights and play chess.
The information acquisition passage of playing chess in the information fusion model of fighting as shown in Figure 1 is respectively automatic vehicle identification system (utilizing the non-contact intelligent IC-card to realize), video camera and weighbridge.Information acquisition passage in the composition model is used for finishing the collection of storing data, vehicle view data, License Plate Image data and vehicle weight information in the vehicle-mounted non-contact intelligent IC-card respectively.Wherein, the identity that radio-frequency (RF) IC card is used for identifying vehicle with relevant simple information is provided, but this sign is not absolutely reliably; Video camera is used for taking vehicle picture and licence plate picture respectively, and weighbridge is responsible for collecting the weight information of vehicle.Key component in the composition model is that the information that collects is carried out information fusion, differentiates the identity of vehicle more accurately with this, thereby guarantees the safety of charge.
In the model, car plate identifying information passage is to obtain the high reliability identifying information of specificity car plate (lack of standardization, stained as shape, block car plate).The system flowchart of car plate identification as shown in Figure 2.The basic process of license plate identification is as follows: when (1) sailed charge station's inlet into when the vehicle of exercising, controller started the video camera that is arranged on vehicle front and side top and takes down vehicle image simultaneously; (2) image of being taken by video camera that contains vehicle license carries out pre-service by image pick-up card input computing machine.The image pre-service comprises image transitions, figure image intensifying, filtering and image level rectification etc.; (3) carry out licence plate search and location by the range searching module, and be partitioned into the rectangular area that comprises the licence plate character number; (4) the licence plate character is carried out binaryzation, inverse is handled, and is unified into white gravoply, with black engraved characters, and the filtering frame also is partitioned into single character, and the input character recognition subsystem is discerned after normalization.
Utilizing template matching method that license plate area is positioned is the core of license plate identification algorithm, and it has made full use of the abundant characteristics in license plate area edge, and accurate positioning has improved overall discrimination.
Vehicle identifying information passage obtains vehicle information based on the vehicle recognition technology of dynamic video in the model.By dynamic image being handled the profile that obtains by vehicle, and with Freeman chain representation profile, obtain vehicle commander, overall height, girth, area, vehicle commander's height ratio, circularity, boundary rectangle and 7 features such as area compares by statistics and calculating, also extracted 7 invariant moment features of vehicle image simultaneously the Freeman chain code.Use the strategy that SVM technology and decision tree multicategory classification device combine then and classify, make sorter have good classification performance and robustness.System flow as shown in Figure 3, model recognition system hardware is made up of ccd video camera, image pick-up card, pattern process computer.Ccd video camera, image pick-up card are responsible for gathering video data, and become data-switching digital picture to send pattern process computer to handle.Pattern process computer has referred to install the computing machine of model recognition system software.It is bearing Video Capture, moving object detection, cuts apart image pre-service, vital tasks such as image characteristics extraction, image classification.
As for the radio-frequency (RF) identification and the passage of weighing that do not stop, because of the technology of comparative maturity is arranged, we have adopted more advanced technology to come to provide this two information sources for model.
As shown in Figure 4, play chess in the Fusion Model fighting, with the functional mode of this model be decomposed into that the player extracts, strategy extracts, payment is extracted, situation assessment and process extract five modules.
(1) player extracts (Player Refinement)
Play chess in the Fusion Model fighting, be considered as the player participating in the sensor multiple information sources of conflict with cooperation.
(2) strategy extracts (Strategy Refinement)
It is to play chess each related in fusion player to determine the set of strategies that it can be adopted to fighting that strategy extracts.
(3) (Payoff Refinement) extracted in payment
When adopting different strategy combinations between the player, the degree difference that systematic uncertainty is reduced, the quantity of information under therefore in theory Different Strategies being made up is as payoff.And in the application of reality, different strategies can cause the variation difference of system performance, promptly with the variation of concrete some parameter in the concrete system payoff as strategy.
(4) situation assessment (Situation Assessment)
Play chess in the Fusion Model fighting, consider the conflict with cooperative process in different situation come down to adopt different strategies to cause by each player, thereby adopting the combination of Different Strategies to be defined as a situation each player, the payoff under the situation is then as the quantitative measurement value of situation assessment.
(5) process is extracted (Process Refinement)
Play chess in the Fusion Model fighting, the process extraction module has directly been finished " executing " function of emerging system.Under the environment of multi-source information conflict, in order to maximize payment effectiveness, the player adopts different strategies, and these strategies are corresponding to the different parameters pattern of sensing system or different collaboration methods etc., thereby the selection of playing chess optimal strategy in the fusion of fighting finally should act on the sensing system, and promptly emerging system is to sensor " executing ".
(6) data base management system (DBMS) (Database management system)
In the data base management system (DBMS) in this model, policy database and payment data storehouse have been defined.Wherein, policy database is used for the different players' of storage system admissible strategy, and the payoff of Different Strategies combination is then deposited in the payment data storehouse.In actual applications, can determine the database of required employing, can not adopt data base management system (DBMS) under the simple situation according to actual conditions, and with the data storage of reality in internal memory; Can all adopt database to be managed to player, situation etc. under the complicated case.
Play chess in the functional mode of fusion whole fighting, determined three's functional module,, determined situation assessment and two modules of process extraction from the purpose that merges from player, strategy, three key elements of payment.In fact, player, strategy and payment three's determines it is prerequisite and the basis that situation assessment and process are extracted.And the whole process that merges is the process of a feedback, and the result that the situation of previous moment is assessed and process is extracted might be to back one player, strategy and the payment three constantly living influence of fixed output quota really.
We have mainly studied under multi-source information conflict environment, playing chess thought to fight serves as to instruct fighting of making up to play chess Fusion Model, inquired into the formalized description of model, the concrete idea about modeling of model, and provided in the general solution fusion process multi-source information conflict and fighting of cooperation and play chess the functional mode of fusion.
Fight and play chess the process of playing chess of fighting alternately that Fusion Model is conflicted multi-source information to carry out with different strategies between the abstract different players that constitute for multisensor of information fusion under the environment, this target of playing chess of fighting is the income of maximization system, just utilizes the information of each sensor substantially.
Fig. 5 is presented as the bayesian network structure algorithm of fighting and playing chess information fusion model of testing vehicle register identification, belongs to the Bayesian inference algorithm in essence.Represent that respectively implication is as follows in the algorithm:
Box indicating variables corresponding (player).Wherein car plate variable (L1) is got character value, and car plate color variance (C1) is quantized (as yellow, blue look etc.), and vehicle variable (T1) is quantized (large-scale, medium-sized, small-sized), and the car weight variable is quantized (as very heavy, heavy, medium, light etc.).L2, T2 variable for reading from radio-frequency (RF) IC card.
P1, p2 ..., pn is one group of parameter of video camera, q1, and q2 ..., qn is one group of parameter of weighing system, they can regard player's a group policy as.
P (L1/ (p1, p2 ..., pn)) for player L1 adopt strategy (p1, p2 ..., the payment in the time of pn), it come down to video camera obtain parameter (p1, p2 ..., under condition pn), license plate identification is the close degree of L1 and actual value as a result.Other P (C1/ (p1, p2 ..., pn)), P (T1/ (p1, p2 ..., pn)) etc. implication identical therewith.
P (L2/L1) takes the payment of corresponding strategy for player L2.The strategy that L2 takes is the value that L2 obtains from the radio-frequency channel.The payment of L2 comes down to obtain the conditional probability that L2 under the condition of certain value obtains certain value at L1.This shows that different players' payment influences each other, show as the mutual fusion of information.Other are identical therewith as P (T2/C1), P (T2/T1), P (T2/W1) implication.
P (J/L2) is for the payment during as the player is subjected to the influence degree of L2 with whole emerging system (J), and the implication of P (J/T2) is identical therewith.The strategy of J has only two: value is legal vehicle or illegal vehicle.The payment difference was correct strategy and pay big strategy when J got Different Strategies.The payment of J is subjected to the influence of L2 and T2, has embodied fusion.
Testing vehicle register identification among the ETC specifically implements, and realizes fighting playing chess information fusion and decision making function with the Bayesian network reasoning algorithm.As can be seen from Figure 5, this Bayesian network algorithm structure is made up of two branches, is respectively car plate identification branch and vehicle identification branch.
One, car plate identification branch
1, algorithm thought
Discern this road information channel for car plate, (bear according to observed reading by machine, compare reading the car plate that license board information and car plate cognitron identify in the IC-card, draw and conform to and be not inconsistent two kinds of observed results), utilize Bayesian formula, releasing car plate cognitron recognition result is the probability distribution of " correctly " and " incorrect ".If push away to such an extent that car plate cognitron recognition result is that correct probability is very big, and incorrect probability is very little, under the condition of conforming to, then system can push away to such an extent that this is " legal vehicle ".If observed result is not for being inconsistent, utilize Bayesian formula, releasing car plate cognitron recognition result is the probability distribution of " correctly " and " incorrect ", if push away to such an extent that license plate recognition result is the probability very big (surpassing a certain thresholding) of " correct identification ", then can be estimated as " illegal vehicle ", if probability very little (less than a certain thresholding) then can not be released the conclusion of " illegal vehicle ".This moment need be with reference to the probability distribution of other information channels.
Algorithm needs some prior probabilities of input when moving for the first time, after this prior probability need constantly obtain according to the operation of algorithm revising becoming posterior probability.
2, arthmetic statement is as shown in Figure 6:
L1: identification correct (getting the a1 value): probability: 95%
Discern incorrect (getting the b1 value): probability: 5%
L2: conform to (c1)/correct identification (a1): conditional probability: 99.9%
Be not inconsistent (d1)/correct identification (a1): conditional probability: 0.1%
Conform to (c1)/not correctly discern (b1): conditional probability: 0.001%
Be not inconsistent (d1)/not and correctly discern (b1): conditional probability: 99.999%
Prior probability is as follows:
p(a1)＝95％
p(b1)＝5％
p(c1/a1)＝99.9％
p(d1/a1)＝0.1％
p(c1/b1)＝0.001％
p(d1/b1)＝99.999％
If observed result is: conform to (meaning is p (c1)=1) or be not inconsistent (meaning is p (d1)=1)
Then demand goes out the license plate recognition result distribution of this Vehicle License Plate Recognition System.That is:
Conform to: p (c1)=1
p(a1/c1)＝p(c1/a1)p(a1)/(p(c1/a1)p(a1)+p(c1/b1)p(b1))
p(b1/c1)＝p(c1/b1)p(b1)/(p(c1/a1)p(a1)+p(c1/b1)p(b1))
Be not inconsistent: p (d1)=1
p(a1/d1)＝p(d1/a1)p(a1)/(p(d1/a1)p(a1)+p(d1/b1)p(b1))
p(b1/d1)＝p(d1/b1)p(b1)/(p(d1/a1)p(a1)+p(d1/b1)p(b1))
Two, vehicle identification branch
Vehicle identification branch is made up of three branches, is respectively the identification of car plate color, vehicle identification and car weight identification.T2 is the vehicle code that IC-card is read, and the algorithm of this branch is the same substantially with car plate identification branch.
Discern for the car plate color, because different colours car plate and vehicle have certain corresponding relation,, become mapping table to show and conform to according to the machine vision result, do not become mapping table to show and be not inconsistent, be not inconsistent the probability distribution that to utilize Bayesian formula to calculate car plate color identification correctness under the condition.Identification for vehicle, because we only recognize large-scale, medium-sized, small-sized three kinds of vehicles, therefore a corresponding relation with the vehicle code that reads from IC-card is also arranged, setting up expression when corresponding relation conforms to, the expression of being false is not inconsistent, be not inconsistent under the condition, needing to calculate the probability distribution of vehicle identification correctness, identification also makes and uses the same method for car weight.
In each bar branch of vehicle identification, the probability distribution of car plate color identification, the probability distribution of vehicle identification, the probability distribution that car weight is measured are weighted on average, can obtain the probability distribution that whole vehicle is discerned branch.
The multi-source information of playing chess information fusion of being used for fighting comprises vehicle identity information that IC-card stores, car plate identifying information, vehicle identifying information and from the car weight information of the weighing system that do not stop.These information come from camera chain, radio-frequency (RF) IC card read-write system and the portable weighing system that do not stop that is arranged on the lane in which the drivers should pay fees trackside respectively.After the independent respectively collection of these information, mode with serial communication sends lane control machine (generally being served as by industrial computer) to, play chess the information fusion center by fighting on the industrial computer and adopt the Bayesian network reasoning algorithm, these information are fought play chess fusion, finally judge the true identity of vehicle, provide reliable foundation for charge station realizes the automatic charge of ETC.
Software adopts that OO high-level programming language---DELPHI writes among the present invention.Program circuit is seen shown in Figure 7.As one of core of ETC vehicle line toll software, the not independent interface of using of this software, it is to be embedded in the complete vehicle line toll software, and reflects in the vehicle line toll software interface of ETC.What the right half part of the vehicle line toll software interface that we develop reflected is the processing of the testing vehicle register identification among the ETC.
Contain ETC of the present invention system and adopted comparatively advanced radio-frequency recognition system, the radio-frequency (RF) identification among the ETC (RFID) system is called automotive vehicle identification (AVI) system again, is by to vehicle-mounted radio-frequency (RF) IC card read-write identification vehicle, finishes charge.Program flow diagram (comprising initial phase) as shown in Figure 8.
Claims (5)
1, the vehicle personal identification system in the ETC electric non-stop toll, comprise the vehicle-mounted radio-frequency (RF) IC card and the read-write system thereof that there are vehicle essential information data, it is characterized in that containing a plurality of information channels that are connected with a plurality of sensors, each information channel is played chess information fusion model with fighting to be organized, with the reasoning algorithm of Bayesian network, finish accurate identification to testing vehicle register; Described fighting played chess information fusion model and comprised player, strategy and pay three fundamentals, the player is meant to fight and plays chess the participator of fusion, a plurality of information channels that promptly being used to fights plays chess fusion, strategy be to the player when the various situation that may occur in playing chess is fought in processing, a kind of description of the decision-making of doing, be the mode of operation or the running parameter of each information channel, payment is the degree of confidence of each information channel to the decision support of information fusion center.
2, the vehicle personal identification system in the ETC electric non-stop toll as claimed in claim 1, it is characterized in that: described fight play chess that information fusion model comprises that the player extracts, strategy extracts, payment is extracted, situation assessment, process extract five modules and data base management system (DBMS), described data base management system (DBMS) is made up of policy database and payment data storehouse, wherein policy database is used for the different players' of storage system admissible strategy, and the payoff of Different Strategies combination is then deposited in the payment data storehouse; Described reasoning algorithm is non-tree type Bayesian network reasoning algorithm.
3, the vehicle personal identification system in the ETC electric non-stop toll as claimed in claim 1 or 2 is characterized in that: the described a plurality of information channels that are used to merge comprise license board information and vehicle information, and described sensor is video camera and weighing system.
4, the vehicle personal identification system in the ETC electric non-stop toll as claimed in claim 3, it is characterized in that: described vehicle information comprises car plate color, vehicle and car weight, wherein the car plate color comprises yellow, blue look, black, white, it is large-scale, medium-sized, small-sized that vehicle comprises, that car weight comprises is very heavy, heavy, medium, gently, very light five quantized levels; License board information and car plate color, vehicle are all gathered as sensor by video camera, and car weight is gathered as sensor by weighing system.
5, the vehicle personal identification system in the ETC electric non-stop toll as claimed in claim 4 is characterized in that: the license board information matching degree in described license plate image identification channel information and radio-frequency (RF) identification passage is big more, and this passage payment is big more; Vehicle information conforms degree in vehicle image recognition channel information and the radio-frequency (RF) identification passage is big more, and this passage payment is big more, but under identical payment, big not as good as license plate image identification channel information to the influence of emerging system decision-making; Vehicle weight information and car plate colouring information are all relevant with vehicle information in the radio-frequency (RF) identification passage, determine their payment size according to the size of correlativity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2006100180461A CN101097605B (en) | 2006-06-30 | 2006-06-30 | Vehicle personal identification system of ETC electric no-parking charge |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2006100180461A CN101097605B (en) | 2006-06-30 | 2006-06-30 | Vehicle personal identification system of ETC electric no-parking charge |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101097605A true CN101097605A (en) | 2008-01-02 |
| CN101097605B CN101097605B (en) | 2010-12-29 |
Family
ID=39011430
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2006100180461A Expired - Fee Related CN101097605B (en) | 2006-06-30 | 2006-06-30 | Vehicle personal identification system of ETC electric no-parking charge |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101097605B (en) |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102184412A (en) * | 2011-05-09 | 2011-09-14 | 东南大学 | Licence plate number and letter identification method based on minimum-error-rate Bayes classifier |
| CN102280032A (en) * | 2010-06-13 | 2011-12-14 | 中兴通讯股份有限公司 | Vehicle traffic management system, method and device |
| CN102467757A (en) * | 2010-11-12 | 2012-05-23 | 艾伯资讯（深圳）有限公司 | Multifunctional electronic toll collection on-board unit |
| CN102034279B (en) * | 2009-09-25 | 2013-01-16 | 无锡德飞科技有限公司 | Bill drawing and checking method for parking place |
| CN103065367A (en) * | 2012-12-17 | 2013-04-24 | 山东易构软件技术有限公司 | Barrier-free dynamic weighing system for road toll |
| CN103150769A (en) * | 2013-01-05 | 2013-06-12 | 山西国强科技发展有限责任公司 | Lorry weighing overload governance and non-stop electronic toll collection two-in-one system device |
| CN103177478A (en) * | 2013-04-09 | 2013-06-26 | 杨红芳 | Data processing method and device based on image identification |
| CN103324618A (en) * | 2012-03-20 | 2013-09-25 | 澳门大学 | Video detection method and system based on electronic imprint |
| CN104134239A (en) * | 2014-07-28 | 2014-11-05 | 中国科学院自动化研究所 | System and method for improving electronic charging reliability in city intelligent transportation system |
| TWI470586B (en) * | 2011-09-29 | 2015-01-21 | Mitsubishi Heavy Ind Ltd | Apparatus for measuring object, method for measuring object, and program therefor |
| TWI483221B (en) * | 2011-09-29 | 2015-05-01 | Mitsubishi Heavy Ind Ltd | Apparatus for measuring object, and processing method and program therefor |
| CN104809763A (en) * | 2015-04-30 | 2015-07-29 | 深圳市车易泊技术股份有限公司 | Parking charging and payment systems and parking charging and payment methods thereof |
| CN105513213A (en) * | 2016-02-04 | 2016-04-20 | 李勇妹 | Electric automobile charging pile capable of identifying identity in unmanned charging process |
| CN105574485A (en) * | 2015-11-11 | 2016-05-11 | 北京四象网讯科技有限公司 | Vehicle information identification method and system |
| CN105574944A (en) * | 2015-12-15 | 2016-05-11 | 重庆凯泽科技有限公司 | Highway intelligent toll collection system based on vehicle identification and method thereof |
| CN105678910A (en) * | 2016-02-04 | 2016-06-15 | 李小春 | Dual-mode charging pile |
| CN105678911A (en) * | 2016-02-04 | 2016-06-15 | 彭晓梅 | Electric automobile charging pile with WIFI hot spot charging function |
| CN105761362A (en) * | 2016-02-04 | 2016-07-13 | 常英梅 | APP payment system applied to electric vehicle charging pile |
| CN105809752A (en) * | 2014-12-30 | 2016-07-27 | 航天信息股份有限公司 | ETC system integrated with infrared camera function, and vehicle identification and location method |
| CN106441132A (en) * | 2016-10-31 | 2017-02-22 | 广东工业大学 | Optical and mechanical mixed multipoint projection self-calibration aperture measuring system and method for nonstandard inner hole |
| CN107248286A (en) * | 2017-06-08 | 2017-10-13 | 安徽和力成信息科技有限公司 | A kind of remote supervision system of high speed ETC tracks operation |
| CN107767472A (en) * | 2016-08-15 | 2018-03-06 | 天津科畅慧通信息技术有限公司 | A kind of system for solving expressway tol lcollection mouth failure |
| CN109615849A (en) * | 2019-01-08 | 2019-04-12 | 河北省高速公路青银管理处 | A kind of new-type monitoring method for running and system of super expressway |
| CN110378658A (en) * | 2019-07-26 | 2019-10-25 | 深圳成谷科技有限公司 | Vehicle overload checks service system, method, lane, auxiliary charging device |
| CN111340966A (en) * | 2019-12-12 | 2020-06-26 | 山东中创软件工程股份有限公司 | Vehicle detection method, device and system |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5933096A (en) * | 1997-01-17 | 1999-08-03 | Mitsubishi Denki Kabushiki Kaisha | Non-stop automatic toll collection system |
| CN1253833C (en) * | 2002-04-25 | 2006-04-26 | 深圳市深港产学研数码科技有限公司 | Method for charging moving vehicle on road and system thereof |
| AT414280B (en) * | 2002-09-12 | 2006-11-15 | Siemens Ag Oesterreich | METHOD FOR IDENTIFYING A MAJOR ROAD SECTION |
| JP2004227185A (en) * | 2003-01-21 | 2004-08-12 | Toshiba Corp | Antenna unit and card processing system |
| CN1707525A (en) * | 2005-05-08 | 2005-12-14 | 广东联合电子收费股份有限公司 | Combined road bridge automatic fees-collecting method based on inductive vehicular unit |
-
2006
- 2006-06-30 CN CN2006100180461A patent/CN101097605B/en not_active Expired - Fee Related
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102034279B (en) * | 2009-09-25 | 2013-01-16 | 无锡德飞科技有限公司 | Bill drawing and checking method for parking place |
| CN102280032A (en) * | 2010-06-13 | 2011-12-14 | 中兴通讯股份有限公司 | Vehicle traffic management system, method and device |
| CN102280032B (en) * | 2010-06-13 | 2014-01-01 | 中兴通讯股份有限公司 | Vehicle traffic management system, method and device |
| CN102467757A (en) * | 2010-11-12 | 2012-05-23 | 艾伯资讯（深圳）有限公司 | Multifunctional electronic toll collection on-board unit |
| CN102184412B (en) * | 2011-05-09 | 2013-05-01 | 东南大学 | Licence plate number and letter identification method based on minimum-error-rate Bayes classifier |
| CN102184412A (en) * | 2011-05-09 | 2011-09-14 | 东南大学 | Licence plate number and letter identification method based on minimum-error-rate Bayes classifier |
| TWI483221B (en) * | 2011-09-29 | 2015-05-01 | Mitsubishi Heavy Ind Ltd | Apparatus for measuring object, and processing method and program therefor |
| TWI470586B (en) * | 2011-09-29 | 2015-01-21 | Mitsubishi Heavy Ind Ltd | Apparatus for measuring object, method for measuring object, and program therefor |
| CN103324618A (en) * | 2012-03-20 | 2013-09-25 | 澳门大学 | Video detection method and system based on electronic imprint |
| CN103324618B (en) * | 2012-03-20 | 2016-08-03 | 澳门大学 | A kind of video detecting method based on electronics trace and system |
| CN103065367A (en) * | 2012-12-17 | 2013-04-24 | 山东易构软件技术有限公司 | Barrier-free dynamic weighing system for road toll |
| CN103150769A (en) * | 2013-01-05 | 2013-06-12 | 山西国强科技发展有限责任公司 | Lorry weighing overload governance and non-stop electronic toll collection two-in-one system device |
| CN103177478A (en) * | 2013-04-09 | 2013-06-26 | 杨红芳 | Data processing method and device based on image identification |
| CN104134239A (en) * | 2014-07-28 | 2014-11-05 | 中国科学院自动化研究所 | System and method for improving electronic charging reliability in city intelligent transportation system |
| CN105809752A (en) * | 2014-12-30 | 2016-07-27 | 航天信息股份有限公司 | ETC system integrated with infrared camera function, and vehicle identification and location method |
| CN104809763A (en) * | 2015-04-30 | 2015-07-29 | 深圳市车易泊技术股份有限公司 | Parking charging and payment systems and parking charging and payment methods thereof |
| CN105574485A (en) * | 2015-11-11 | 2016-05-11 | 北京四象网讯科技有限公司 | Vehicle information identification method and system |
| CN105574944A (en) * | 2015-12-15 | 2016-05-11 | 重庆凯泽科技有限公司 | Highway intelligent toll collection system based on vehicle identification and method thereof |
| CN105678910A (en) * | 2016-02-04 | 2016-06-15 | 李小春 | Dual-mode charging pile |
| CN105761362A (en) * | 2016-02-04 | 2016-07-13 | 常英梅 | APP payment system applied to electric vehicle charging pile |
| CN105678911A (en) * | 2016-02-04 | 2016-06-15 | 彭晓梅 | Electric automobile charging pile with WIFI hot spot charging function |
| CN105513213A (en) * | 2016-02-04 | 2016-04-20 | 李勇妹 | Electric automobile charging pile capable of identifying identity in unmanned charging process |
| CN105761362B (en) * | 2016-02-04 | 2016-11-30 | 国网山东省电力公司费县供电公司 | It is applied to the APP payment system of charging pile for electric vehicle |
| CN107767472A (en) * | 2016-08-15 | 2018-03-06 | 天津科畅慧通信息技术有限公司 | A kind of system for solving expressway tol lcollection mouth failure |
| CN106441132A (en) * | 2016-10-31 | 2017-02-22 | 广东工业大学 | Optical and mechanical mixed multipoint projection self-calibration aperture measuring system and method for nonstandard inner hole |
| CN107248286A (en) * | 2017-06-08 | 2017-10-13 | 安徽和力成信息科技有限公司 | A kind of remote supervision system of high speed ETC tracks operation |
| CN109615849A (en) * | 2019-01-08 | 2019-04-12 | 河北省高速公路青银管理处 | A kind of new-type monitoring method for running and system of super expressway |
| CN110378658A (en) * | 2019-07-26 | 2019-10-25 | 深圳成谷科技有限公司 | Vehicle overload checks service system, method, lane, auxiliary charging device |
| CN111340966A (en) * | 2019-12-12 | 2020-06-26 | 山东中创软件工程股份有限公司 | Vehicle detection method, device and system |
| CN111340966B (en) * | 2019-12-12 | 2022-03-08 | 山东中创软件工程股份有限公司 | Vehicle detection method, device and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101097605B (en) | 2010-12-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101097605B (en) | Vehicle personal identification system of ETC electric no-parking charge | |
| CN106127107A (en) | The model recognizing method that multi-channel video information based on license board information and vehicle's contour merges | |
| CN106910369B (en) | Roadside Parking management method and system | |
| CN109035447A (en) | A kind of Logistics Park vehicle parking charging method and system | |
| CN110310378B (en) | Open parking lot parking charging method and system based on double technologies | |
| CN109033175A (en) | A kind of method and system to scheme to search vehicle | |
| CN103971097A (en) | Vehicle license plate recognition method and system based on multiscale stroke models | |
| CN107153933A (en) | A kind of vehicle management system based on RFID | |
| CN100343856C (en) | Vehicle renting management system and method | |
| CN109670449B (en) | Vehicle illegal judgment method based on vertical snapshot mode | |
| CN105404858A (en) | Vehicle type recognition method based on deep Fisher network | |
| CN110232827A (en) | The recognition methods of free flow toll vehicle type, apparatus and system | |
| CN103679209A (en) | Sparse theory based character recognition method | |
| CN204833376U (en) | System for prevent that vehicle from evading expense | |
| CN1825359A (en) | Prepayment, giving change IC card read-write device, smart card subsystem and paying method | |
| CN1851755A (en) | Prepaid change-giving IC card read-write device, one-card-through subsystem and its realizing method | |
| CN105139462A (en) | Two-dimension-code-based method for preventing vehicle passing through tollgate forcibly on highway | |
| CN1838177A (en) | IC-card reader-writer for prepayment and change, traffic card subsystem and implementing method thereof | |
| CN203012798U (en) | Full-automatic expressway vehicle charging system | |
| CN111417088B (en) | Identity attribute authentication method for Internet of vehicles data | |
| CN204884183U (en) | Parking area bluetooth WIFI read head and cell -phone APP assist license plate recognition | |
| CN211928668U (en) | Bus-mounted POS machine riding system based on face recognition as auxiliary authentication | |
| CN102682299A (en) | Character overall feature-based rapid number plate verification method | |
| CN106347151A (en) | Energy supplementing system | |
| CN110110721A (en) | Vehicle information recognition method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20101229 Termination date: 20130630 |