CN103093621A - Processing method and device of multisource traffic information fusion - Google Patents

Abstract

The invention discloses a processing method and a device of multisource traffic information fusion so as to enable a link region and a license plate range to obtain accurate fusion value, the fusion value of the license plate range can reflect the road condition of long road sections accurately and provide accurate road condition information, and the fusion value of the link region can reflect the road condition of short road sections accurately and provide accurate route guidance. The method comprises the steps that multisource data are obtained; pre-processing is conducted to the multisource data; the multisource data after pre-processing are respectively matched to the license plate region and the link region; fusion models of the multisource data are respectively chosen for the license plate region and the link region according to data deficiency circumstance of the multisource data and a deficiency supplying method adopted by the multisource data; and data fusion is respectively conducted to the multisource data of the license plate region and the link region according to the obtained fusion models, and therefore fusion value of the license plate region and the link region is obtained. The processing method and the device of the multisource traffic information fusion are applied to information fusion technology.

Description

Disposal route and device that a kind of multi-source traffic information merges

Technical field

The present invention relates to information fusion technology, relate in particular to disposal route and device that a kind of multi-source traffic information merges.

Background technology

The through street is as the chief component of city road network, urban transportation is played an important role, therefore, assurance in real time road conditions transport information comprehensively and accurately is the prerequisite that road network is control effectively, and is also the key of alleviating urban traffic blocking, reducing traffic hazard.

In existing traffic system, different checkout equipments there are differences in all many-sides, as detected parameters, coverage, data precision, acquisition cost etc.In the real data gatherer process, due to the data error of extremely bringing of various detecting devices, the inconsistent and deficient phenomena between data, caused the limitation of the information exported in present intelligent transportation system, the shared difficulty of each traffic subsystem information is large.For above problem, can realize by Data fusion technique the verification that complements each other of multi-source traffic information, thereby make the traffic state information of road network more accurate.

Multi-source traffic information refers to the detection data that a plurality of traffic checkout equipments obtain, and a plurality of checkout equipments generally comprise: Floating Car detection system, microwave detector, licence plate detecting device etc.The detection data that the Floating Car detection system obtains are take the link interval as the floating car data of unit, comprising: Floating Car sample size, distance, time; The detection data that microwave detector obtains be take between the microwave region as the microwave data of unit, comprising: flow, speed, time occupancy; The detection data that the licence plate detecting device obtains are take the licence plate interval as the licence plate data of unit, comprising: distance, time.Generally, the licence plate interval comprises between a plurality of microwave detection zones, comprises a plurality of links between each microwave region interval.

The licence plate data that the licence plate detecting device obtains can reflect road conditions accurately, and still, the licence plate detecting device is a kind of in fixed detector, and its distribution on the through street is limited, can not obtain the traffic information of whole through street nets.When the licence plate shortage of data, can pass through Data fusion technique, floating car data and microwave data are merged, obtain road conditions more accurately.

In prior art, only for multi-source data in the licence plate interval or the link interval carry out data fusion.When carrying out data fusion for the licence plate interval, the Fusion Model of employing is not necessarily accurate for the link interval, causes the inaccurate of fusion value, and then causes the path navigation that provides inaccurate; When carrying out data fusion for the link interval, the Fusion Model of employing is not necessarily accurate for the licence plate interval, causes the traffic information that provides inaccurate.And, in prior art, when data fusion is carried out in licence plate interval or link interval, all only adopting a kind of Fusion Model, the fusion value that obtains is not accurate enough.

Summary of the invention

Disposal route and device that embodiments of the invention provide a kind of multi-source traffic information to merge, make the interval and licence plate of link interval can obtain to merge more accurately value, and then the fusion value that makes the licence plate interval can reflect the road conditions in long highway section accurately, traffic information accurately is provided, the fusion value in link interval can reflect the road conditions on short chain road accurately, provides accurately the path to induce.

For achieving the above object, embodiments of the invention adopt following technical scheme:

First aspect the invention provides the disposal route that a kind of multi-source traffic information merges, and the method comprises:

Obtain multi-source data, described multi-source data comprises floating car data, microwave data;

Described multi-source data is carried out pre-service, and described pre-service comprises that conversion is unified, misdata is rejected, data are filled a vacancy and time match;

Described pretreated multi-source data is matched respectively the interval and link of licence plate interval;

According to the shortage of data situation of described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model;

According to the described Fusion Model of choosing, the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.

The treating apparatus that second aspect, the embodiment of the present invention provide a kind of multi-source traffic information to merge, described device comprises: acquiring unit, pretreatment unit, matching unit, choose unit, integrated unit;

Described acquiring unit is used for obtaining multi-source data, and described multi-source data comprises floating car data, microwave data;

Described pretreatment unit, the multi-source data that is used for described acquiring unit is obtained carries out pre-service, and described pre-service comprises that conversion is unified, misdata is rejected, data are filled a vacancy and time match;

Described matching unit, licence plate is interval and link is interval for the pretreated multi-source data of described pretreatment unit is matched respectively;

The described unit of choosing is used for the shortage of data situation according to described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, and the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model;

Described integrated unit is used for according to the described Fusion Model of choosing unit selection, and the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.

Disposal route and device that the embodiment of the present invention provides a kind of multi-source traffic information to merge obtain multi-source data, and described multi-source data is carried out pre-service, will described pretreated multi-source data match respectively the interval and link of licence plate interval; According to the shortage of data situation of described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model; According to the described Fusion Model of choosing, the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.Owing to and link interval to licence plate interval respectively choosing Fusion Model according to the deletion condition of multi-source data and the method for filling a vacancy of utilization, and carry out data fusion according to the Fusion Model of choosing, make the interval and licence plate of link interval can obtain to merge more accurately value, and then the fusion value that makes the licence plate interval can reflect the road conditions in long highway section accurately, traffic information accurately is provided, the fusion value in link interval can reflect the road conditions of shorted segment accurately, provides accurately the path to induce.

Description of drawings

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, the below will do to introduce simply to the accompanying drawing of required use in embodiment or description of the Prior Art, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skills, under the prerequisite of not paying creative work, can also obtain according to these accompanying drawings other accompanying drawing.

The schematic flow sheet of the disposal route that a kind of multi-source traffic information that Fig. 1 provides for the embodiment of the present invention merges;

The schematic flow sheet of the disposal route that the another kind of multi-source traffic information that Fig. 2 provides for the embodiment of the present invention merges;

The structural representation of the treating apparatus that a kind of multi-source traffic information that Fig. 3 provides for the embodiment of the present invention merges.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is only the present invention's part embodiment, rather than whole embodiment.Based on the embodiment in the present invention, those of ordinary skills belong to the scope of protection of the invention not making the every other embodiment that obtains under the creative work prerequisite.

Embodiment one,

The disposal route that the embodiment of the present invention provides a kind of multi-source traffic information to merge, as shown in Figure 1, the method comprises:

101, obtain multi-source data, described multi-source data comprises floating car data, microwave data.

Fixed detector on city expressway comprises microwave detector and licence plate detecting device.Also exist mobile Floating Car for detection of traffic data on city expressway.

Wherein, the traffic data precision that the licence plate detecting device obtains is high, but the negligible amounts of licence plate detecting device, the data volume of acquisition is less, and the transport information that can reflect is limited.But for the highway section that there is no the licence plate detecting device, if only utilize single detecting device, detect or the detection of Floating Car detecting device as microwave, the data of detection can accurately not reflect traffic information.Therefore, carry out the fusion of multi-source traffic information, transport information is necessary more accurately in acquisition.

Wherein, the data volume of the floating car data that the Floating Car detection system detects is more, can reflect transport information in a big way.Therefore, in the embodiment of the present invention, adopt the microwave data of floating car data and microwave detector detection as multi-source data, carry out the fusion of multi-source traffic information, obtain wide coverage, and the higher traffic data of precision.

Wherein, the data of Floating Car detection system acquisition comprise: Floating Car Average Travel Speed and Floating Car sample size.Wherein, the Floating Car Average Travel Speed represents the average overall travel speed in corresponding link interval; The Floating Car sample size represents the quantity of the Floating Car of travelling on corresponding link.

The data that microwave detector obtains comprise: flow, speed and time occupancy.Wherein, flow represents in the unit interval by the vehicle number between a certain microwave region; Speed represents the vehicle instantaneous velocity mean value by detecting device; Time occupancy represents time that microwave detector in specific observation time is taken by vehicle and the ratio of observation time.Time occupancy hour, the vehicle by microwave detector in the unit interval is less, and travel speed is higher, causes time occupancy lower; When time occupancy was larger, the vehicle by microwave detector in the unit interval was more, and the speed of travelling is lower, causes time occupancy higher.

In the disposal route that the multi-source traffic information of the embodiment of the present invention merges, described multi-source data can be specifically floating car data, microwave data.

Concrete, floating car data can obtain by the floating vehicle system collection, and microwave data can obtain by the microwave detector collection.

102, described multi-source data is carried out pre-service, described pre-service comprises that conversion is unified, misdata is rejected, data are filled a vacancy and time match.

After obtaining described multi-source data, need to change unification to described multi-source data, make described multi-source data can be expressed as the data of unified parameter, the data of described unified parameter can be: speed, time occupancy etc. can reflect the parameter of road conditions.The embodiment of the present invention is all unified floating car data, microwave data and licence plate data to characterize with identical parameters such as flow, speed, time occupancies.

Misdata refers to have occured do not meet the traffic parameter data of convention sudden change in certain independent sampling interval, normally because detecting device, transmission line failure and vehicles failed cause by the traffic control reason such as travel.When traffic detecting device or transmission line broke down, the data that collect were normally wrong, can not reflect real traffic.The embodiment of the present invention is rejected after adopting threshold method and traffic flow theory to judge to the processing of misdata.

Adopting threshold method and traffic flow theory that misdata is rejected, is technology well known to those skilled in the art, and the embodiment of the present invention no longer specifically describes at this.

Shortage of data can be divided into two classes: a class is the inherent data disappearance, this is because the detecting device sweep frequency is fixing, equipment breaks down, the excessive intensive detecting device that causes of vehicle can't normally detect the reasons such as vehicle, the dynamic data that collects can't be uploaded in strict accordance with the fixed time interval, the shortage of data that causes; Another kind of is the shortage of data that causes after misdata is rejected.

When shortage of data, can fill a vacancy to multi-source data, the concrete method of filling a vacancy comprises: time series method, historical data base method and the spatial position data method of filling a vacancy of filling a vacancy of filling a vacancy.

Because each checkout equipment works alone, the interval of the data that each checkout equipment detects is also not necessarily identical, therefore, need to carry out time match to described multi-source data.The time match of multi-source data refers to that the multi-source traffic flow data that will collect mates, and makes the transport information of data in the same time period of reflection.

Multi-source data is carried out time match, is technology well known to those skilled in the art, and the embodiment of the present invention is not described specifically at this.

103, described pretreated multi-source data is matched respectively licence plate interval and link interval.

The licence plate interval is the fundamental space unit of the data of licence plate detecting device detection, is the highway section with directivity that forms after according to certain rule, road being divided.

The link interval is the fundamental space unit of the data of Floating Car detection system detection, is the highway section with directivity that forms after according to certain rule, road being divided.

The fundamental space unit of the data that between the microwave region, microwave detector detects is the highway sections with directivity that form after according to certain rule, road being divided.

Need to prove, generally, the licence plate interval comprises between a plurality of microwave regions, and comprises between each microwave region that a plurality of links are interval.

Therefore the traffic information that can accurately reflect long highway section due to the data in licence plate interval, in the multi-source traffic information integration technology, matches the licence plate interval with the multi-source data that obtains, and makes multi-source data can reflect accurately that the road conditions in long highway section are necessary.

The traffic information that can reflect accurately shorted segment due to the data in link interval, therefore, in the multi-source traffic information integration technology, the multi-source data that obtains is matched the link interval, make multi-source data can reflect accurately the road conditions of shorted segment, can further provide accurately the path to induce.

In the present embodiment, in the method for amalgamation processing of multi-source traffic information, will carry out pretreated multi-source data and match respectively the interval and link of licence plate interval.

104, according to the shortage of data situation of described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model.

The method of filling a vacancy that adopts when filling a vacancy according to the deletion condition of described multi-source data and data is for the multi-source data in licence plate interval is chosen the interval Fusion Model of licence plate.

The method of filling a vacancy that adopts when filling a vacancy according to the deletion condition of described multi-source data and data is for the multi-source data in link interval is chosen the interval Fusion Model of link.

105, according to the described Fusion Model of choosing, the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.

The Fusion Model of choosing according to the licence plate interval, Floating Car Average Travel Speed to described licence plate in the multi-source data in interval and link interval and the flow in microwave detection data, speed, time occupancy are as the input of Fusion Model, carry out data fusion, obtain the fusion value in the interval and link of licence plate interval.

The disposal route that the embodiment of the present invention provides a kind of multi-source traffic information to merge is obtained multi-source data, and described multi-source data is carried out pre-service, will described pretreated multi-source data matches respectively the interval and link of licence plate interval; According to the shortage of data situation of described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model; According to the described Fusion Model of choosing, the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.Owing to and link interval to licence plate interval respectively choosing Fusion Model according to the deletion condition of multi-source data and the method for filling a vacancy of utilization, and carry out data fusion according to the Fusion Model of choosing, make the interval and licence plate of link interval can obtain to merge more accurately value, and then the fusion value that makes the licence plate interval can reflect the road conditions in long highway section accurately, traffic information accurately is provided, the fusion value in link interval can reflect the road conditions of shorted segment accurately, provides accurately the path to induce.

Embodiment two,

The disposal route that the embodiment of the present invention provides a kind of multi-source traffic information to merge, as shown in Figure 2, the method comprises:

201, according to the deletion condition of the first multi-source data and the method for filling a vacancy accordingly, be respectively interval definite corresponding the first Fusion Model of licence plate interval and link.

Before carrying out multisource data fusion, at first need the shortage of data situation that may occur according to multi-source data and the method for filling a vacancy used, obtain many groups of the first multi-source datas of the method for filling a vacancy of every kind of multi-source data deletion condition and correspondence, and according to described many groups the first multi-source data, be deletion condition and the corresponding method of the filling a vacancy training Fusion Model that adopts of every kind of multi-source data.

Described the first multi-source data comprises: the first floating car data and the first microwave data.The deletion condition of described the first multi-source data is for lacking or not lacking.The described method of filling a vacancy comprises: time series method, historical data base method and the spatial position data method of filling a vacancy of filling a vacancy of filling a vacancy.

For deletion condition and the corresponding method of filling a vacancy of every kind of multi-source data, according to the deletion condition of the first multi-source data and the method for filling a vacancy accordingly, be respectively interval definite corresponding the first Fusion Model of licence plate interval and link.

For instance, when the floating car data in the first multi-source data and microwave data do not lack, be respectively interval and interval definite the first Fusion Model of link of licence plate.

202, choose many groups the first multi-source data for described the first Fusion Model.

Choose many groups many group first multi-source datas corresponding with described the first Fusion Model.For example, when the floating car data in the first multi-source data corresponding to described the first Fusion Model and microwave data do not lack, the first multi-source data when choosing many group Floating Car and microwave data and fill a vacancy for described the first Fusion Model.

203, described many groups the first multi-source data is carried out pre-service.

Described pre-service comprises that conversion is unified, misdata is rejected, data are filled a vacancy and time match.

To described many groups the first multi-source data carry out describing in pretreated method and the present embodiment that multi-source data is carried out pretreated method is identical, the embodiment of the present invention does not repeat them here.

204, described pretreated many groups the first multi-source data is matched respectively the licence plate interval or link is interval.

The method that pretreated multi-source data is matched respectively licence plate interval or link interval of describing in the method that described pretreated many groups the first multi-source data is matched respectively licence plate interval or link interval and the present embodiment is identical, and the embodiment of the present invention does not repeat them here.

205, according to described many groups the first multi-source datas, described the first Fusion Model is carried out model training, obtain a plurality of second Fusion Model in the interval and link of described licence plate interval and the fusion value of correspondence.

Utilize described many groups the first multi-source data to carry out model training to described the first Fusion Model, can obtain a plurality of second Fusion Model in the interval and link of licence plate interval and the fusion value of correspondence.

Before carrying out model training, need to utilize data anastomosing algorithm to carry out the structure of Fusion Model.The data anastomosing algorithm that adopts in the present invention is the algorithm of genetic algorithm and backpropagation BP (Back Propagation) neural network algorithm combination.

Wherein, the search procedure of genetic algorithm be a point set from the space to the search of another one point set, be actually a kind of parallel search, be fit to large-scale parallel and calculate.Genetic algorithm is applicable to global search, is not subjected to the constraint of search volume, does not require continuity, can with very large probability from discrete, multipole value, contain noisy higher-dimension problem and find globally optimal solution.

The BP neural network is a kind of Multi-layered Feedforward Networks by the Back Propagation Algorithm training, is one of neural network algorithm that is most widely used at present.The BP neural network model is being most widely used aspect the fusion of processing multi-source traffic data, has very strong Nonlinear Processing ability, and have advantages such as self study, self-organization, concurrency and fault-tolerance, yet, the BP neural network algorithm is comparatively difficult when seeking globally optimal solution, tends to be absorbed in locally optimal solution in solution procedure.

The genetic algorithm that the present invention adopts and the algorithm of BP neural network algorithm combination are can search for the characteristic of globally optimal solution by genetic algorithm, weights and threshold value to neural network are optimized, obtain optimum network weight and the initial value of threshold value, can effectively avoid neural network to be absorbed in the problem of locally optimal solution.When the algorithm of genetic algorithm and the combination of BP neural network algorithm builds Fusion Model, comprise that mainly genetic Algorithm Design, the design of mode input output parameter, the neural network number of plies design, transport function is chosen and the design of hidden layer neuron quantity.

206, obtain described licence plate interval or link interval with as the average relative error of the licence plate data of true value and the fusion value of least error sum-of-squares difference minimum, second Fusion Model corresponding to fusion value of the described difference minimum in described licence plate interval or link interval is defined as the Fusion Model of the multi-source data in described licence plate interval or link interval.

Calculate the average relative error of a plurality of fusion values corresponding to described true value and described a plurality of the second Fusion Model, obtain fusion value corresponding to described a plurality of the second Fusion Model and the difference of described true value, determine the precision of described a plurality of the second Fusion Model.

Calculate the least error quadratic sum of a plurality of fusion values corresponding to described true value and described a plurality of the second Fusion Model, obtain fusion value corresponding to described a plurality of the second Fusion Model and the difference of described true value, determine the validity of described a plurality of the second Fusion Model.

Calculate the least error quadratic sum of a plurality of fusion values corresponding to described true value and described a plurality of the second Fusion Model, obtain fusion value corresponding to described a plurality of the second Fusion Model and the difference of described true value, determine that the validity of described a plurality of the second Fusion Model specifically comprises: calculate without the first multi-source data of data fusion and the least error quadratic sum between described true value; If the least error quadratic sum between described fusion value and described true value is less than the least error between the floating car data in described the first multi-source data or microwave data and described true value square, described the second Fusion Model is effective.

Determine effective the second model in a plurality of the second Fusion Model, determine that a plurality of second the second the highest Fusion Model of Fusion Model precision are the Fusion Model in described licence plate interval or link interval.

More than for obtain the process of Fusion Model for the method for filling a vacancy of a kind of deletion condition of the first multi-source data and utilization, the method of filling a vacancy of other deletion conditions of the first multi-source data and corresponding utilization is obtained the process of Fusion Model, consistent with said process, the embodiment of the present invention does not repeat them here.

By said method, can be for licence plate all situations of the deletion condition of multi-source data in interval and link interval and the utilizable method of filling a vacancy obtain the Fusion Model of correspondence.

207, obtain multi-source data, described multi-source data comprises floating car data, microwave data.

Fixed detector on city expressway comprises microwave detector and licence plate detecting device.Also exist mobile Floating Car for detection of traffic data on city expressway.

Wherein, the traffic data precision that the licence plate detecting device obtains is high, but the negligible amounts of licence plate detecting device, the data volume of acquisition is less, and the transport information that can reflect is limited.But for the highway section that there is no the licence plate detecting device, if only utilize single detecting device, detect or the detection of Floating Car detecting device as microwave, the data of detection can accurately not reflect traffic information.Therefore, carry out the fusion of multi-source traffic information, transport information is necessary more accurately in acquisition.

Wherein, the data volume of the floating car data that the Floating Car detection system detects is more, can reflect transport information in a big way.Therefore, in the embodiment of the present invention, adopt the microwave data of floating car data and microwave detector detection as multi-source data, carry out the fusion of multi-source traffic information, obtain wide coverage, and the higher traffic data of precision.

Wherein, the data of Floating Car detection system acquisition comprise: Floating Car Average Travel Speed and Floating Car sample size.Wherein, the Floating Car Average Travel Speed represents the average overall travel speed in corresponding link interval; The Floating Car sample size represents the quantity of the Floating Car of travelling on corresponding link.

The data that microwave detector obtains comprise: flow, speed and time occupancy.Wherein, flow represents in the unit interval by the vehicle number between a certain microwave region; Speed represents the vehicle instantaneous velocity mean value by detecting device; Time occupancy represents time that microwave detector in specific observation time is taken by vehicle and the ratio of observation time.Time occupancy hour, the vehicle by microwave detector in the unit interval is less, and travel speed is higher, causes time occupancy lower; When time occupancy was larger, the vehicle by microwave detector in the unit interval was more, and the speed of travelling is lower, causes time occupancy higher.

Concrete, floating car data can obtain by the floating vehicle system collection, and microwave data can obtain by the microwave detector collection.

208, described multi-source data is carried out pre-service, described pre-service comprises that conversion is unified, misdata is rejected, data are filled a vacancy and time match.

After obtaining described multi-source data, need to change unification to described multi-source data, make described multi-source data can be expressed as the data of unified parameter, the data of described unified parameter can be: speed, time occupancy etc. can reflect the parameter of road conditions.The embodiment of the present invention is all unified floating car data, microwave data and licence plate data to characterize with identical parameters such as flow, speed, time occupancies.

Change after reunification, described floating car data comprises: link number, date, timestamp, interval travelling speed.As shown in table 1, be the data layout table of floating car data.Wherein, link number refer to Floating Car current the numbering of link of process; Timestamp refers to 00:00 → 24:00 is divided into 720 2min time periods, and numbering is from 1 → 720; Interval travelling speed refers to that all are through the mean value of the link average speed of the Floating Car of current link in each 2min, and unit is km/h.

The data layout table of table 1 floating car data

Link number	Date	Timestamp	Interval travelling speed
				595672002697	20070911	1	46
595672002697	20070911	2	46
				595672002697	20070911	3	34.5
595672002697	20070911	4	34.5
				595672002697	20070911	5	41.4
…	…	…	…
				…	…	…	…
595672002697	20070911		38
				595672002697	20070911		34.5
595672002697	20070911		51.75

Described microwave data comprises: detecting device numbering, date, timestamp, flow, speed, time occupancy.As shown in table 2, be the data layout table of microwave detection data.Wherein, timestamp refers to 00:00 → 24:00 is divided into 720 2min time periods, and numbering is from 1 → 720; Flow (q) refers to the every two minutes vehicle numbers by detecting device; Speed (v) refers to pass through in 2min the vehicle instantaneous velocity mean value of detecting device, unit: km/h; Occupation rate (b) is time occupancy, refers to that the 2min internal detector detects the shared ratio of time of vehicle, span 0～100%.

The data layout table of table 2 microwave data

Misdata refers to have occured do not meet the traffic parameter data of convention sudden change in certain independent sampling interval, normally because detecting device, transmission line failure and vehicles failed cause by the traffic control reason such as travel.When traffic detecting device or transmission line broke down, the data that collect were normally wrong, can not reflect real traffic.The embodiment of the present invention is rejected after adopting threshold method and traffic flow theory to judge to the processing of misdata.

Adopting threshold method and traffic flow theory that misdata is rejected, is technology well known to those skilled in the art, and the embodiment of the present invention no longer specifically describes at this.

Shortage of data can be divided into two classes: a class is the inherent data disappearance, this is because the detecting device sweep frequency is fixing, equipment breaks down, the excessive intensive detecting device that causes of vehicle can't normally detect the reasons such as vehicle, the dynamic data that collects can't be uploaded in strict accordance with the fixed time interval, the shortage of data that causes; Another kind of is the shortage of data that causes after misdata is rejected.

For the situation of shortage of data, according to the method for filling a vacancy, missing data is filled a vacancy.The described method of filling a vacancy comprises: time series method, historical data base method and the spatial position data method of filling a vacancy of filling a vacancy of filling a vacancy.

In prior art, adopt time series the fill a vacancy order of method of method-historical data base of filling a vacancy that the floating car data of disappearance is carried out data and fills a vacancy for floating car data more; Adopt time series method-spatial position data fill a vacancy order of method of method-historical data base of filling a vacancy of filling a vacancy that the microwave data of disappearance is carried out data and fills a vacancy for microwave data more.

In the embodiment of the present invention, in licence plate interval and the interval fusion value more accurately that obtains of link, be that different shortage of data situations is selected the different methods of filling a vacancy for respectively.

When floating car data lacked, the type of the Linktype corresponding according to missing data and the adjacent link of described link was determined the method for filling a vacancy of missing data.

Concrete, as shown in table 3, be the floating car data method table of filling a vacancy.Wherein, " 1_2_3 " expression method of filling a vacancy is sequentially: time series method-spatial position data method-historical data base method of filling a vacancy of filling a vacancy of filling a vacancy; The fill a vacancy order of method of " 1_3 " expression is: time series method-historical data base method of filling a vacancy of filling a vacancy.For instance, when the missing data respective links of described floating car data is positioned on normal highway section, and when its upstream and downstream link all is the highway section, described missing data is carried out data when filling a vacancy, at first utilizing the time series method of filling a vacancy to carry out data to missing data fills a vacancy, fill a vacancy method when inapplicable when the time sequence, utilize the spatial position data method of filling a vacancy to carry out data to missing data and fill a vacancy; When adopting the spatial position data method of filling a vacancy to carry out data when filling a vacancy to missing data, the velocity amplitude of described missing data respective links is the upstream and downstream link average speed.

The table 3 floating car data method table of filling a vacancy

When microwave data lacked, the microwave detector corresponding according to missing data be correlativity ρ 1 and the ρ 2 of two microwave detectors adjacent with upstream and downstream respectively, determined the method for filling a vacancy of the multi-source data of missing data;

Wherein, $\mathrm{\ρ}1,\mathrm{\ρ}2=\frac{\mathrm{cov}(x,y)}{\sqrt{D\left(x\right)}\sqrt{D\left(y\right)}}=\frac{E((x-E\left(x\right))*(y-E\left(y\right)))}{\sqrt{D\left(x\right)}\sqrt{D\left(y\right)}};$

For p1, x, y represent that respectively microwave detector and upstream microwave detector that in historical data base, missing data is corresponding detect the same time sequence traffic information data vector that obtains; For p2, x, y represent that respectively microwave detector and downstream microwave detecting device that in historical data base, missing data is corresponding detect the same time sequence traffic information data vector that obtains.

Concrete, as shown in table 4, be the microwave data method table of filling a vacancy.Wherein, " 1_2_3 " expression method of filling a vacancy is sequentially: time series method-spatial position data method-historical data base method of filling a vacancy of filling a vacancy of filling a vacancy; The fill a vacancy order of method of " 1_3 " expression is: time series method-historical data base method of filling a vacancy of filling a vacancy.For instance, when link corresponding to the missing data of microwave data is positioned on normal highway section, and the correlativity of the link that missing data is corresponding and upstream and downstream link is all greater than 0.7 o'clock, when described missing data is filled a vacancy, at first adopt the time series method of filling a vacancy that missing data is filled a vacancy, fill a vacancy method when inapplicable when the time sequence, adopt the spatial position data method of filling a vacancy that missing data is filled a vacancy; When adopting the spatial position data method of filling a vacancy that missing data is filled a vacancy, the velocity amplitude of the link that described missing data is corresponding is the average velocity of upstream and downstream link.

The table 4 microwave data method table of filling a vacancy

Because each checkout equipment works alone, the interval of the data that each checkout equipment detects is also not necessarily identical, therefore, need to carry out time match to described multi-source data.The time match of multi-source data refers to that the multi-source traffic flow data that will collect mates, and makes the transport information of data in the same time period of reflection.

Multi-source data is carried out time match, is technology well known to those skilled in the art, and the embodiment of the present invention is not described specifically at this.

209, described pretreated multi-source data is matched respectively licence plate interval and link interval.

The licence plate interval is the fundamental space unit of the data of licence plate detecting device detection, is the highway section with directivity that forms after according to certain rule, road being divided.The link interval is the fundamental space unit of the data of Floating Car detection system detection, is the highway section with directivity that forms after according to certain rule, road being divided.The fundamental space unit of the data that between the microwave region, microwave detector detects is the highway sections with directivity that form after according to certain rule, road being divided.

Need to prove, generally, the licence plate interval comprises between a plurality of microwave regions, and comprises between each microwave region that a plurality of links are interval.

Therefore the traffic information that can accurately reflect long highway section due to the data in licence plate interval, in the multi-source traffic information integration technology, matches the licence plate interval with the multi-source data that obtains, and makes multi-source data can reflect accurately that the road conditions in long highway section are necessary.

The traffic information that can reflect accurately shorted segment due to the data in link interval, therefore, in the multi-source traffic information integration technology, the multi-source data that obtains is matched the link interval, make multi-source data can reflect accurately the road conditions of shorted segment, can further provide accurately the path to induce.

The method that floating car data and microwave data is matched the licence plate interval comprises: at first, floating car data is converted to link zone-to-zone travel stream parameter, microwave data is converted to traffic flow parameter between the microwave region; Obtain respectively between link interval, microwave region and account for licence plate length of an interval degree ratio; According to formula (2), described floating car data and microwave data are matched the licence plate interval.Wherein, formula (2) is:

$A_{\mathrm{PL}}=\frac{x_1}{L}\×A_1+\frac{x_2}{L}\×A_2+\·\·\·+\frac{x_i}{L}\×A_i;$ Formula (2)

In formula (2), the i indication board is according to interval i the microwave detector interval that comprises or link; Traffic flow parameter after the APL indication board mates according to interval space can be speed, the volume of traffic and occupation rate; The Ai indication board is according to the traffic flow parameter of interval interior i microwave detector interval or link detecting; The xi indication board is according to the length (m) of interval interior i microwave detector interval or link; The L indication board is according to burst length (m).Floating car data and microwave data are matched the link interval.

The floating car data that the Floating Car detection system obtains is take the link interval as unit, therefore, when multi-source data being matched link when interval, do not need floating car data is mated.Microwave data is matched link when interval, owing to comprising between the microwave region that link is interval, therefore, the data of the microwave data between the microwave region as the link interval that comprises between the microwave region are got final product.Link for crossing over two microwave detectors mates according to formula (2).

210, according to the shortage of data situation of described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model.

The method of filling a vacancy that adopts when filling a vacancy according to the deletion condition of described multi-source data and data is for the multi-source data in licence plate interval is chosen the interval Fusion Model of licence plate.The method of filling a vacancy that adopts when filling a vacancy according to the deletion condition of described multi-source data and data is for the multi-source data in link interval is chosen the interval Fusion Model of link.

211, according to the described Fusion Model of choosing, the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.

The Fusion Model of choosing according to the licence plate interval, Floating Car Average Travel Speed to described licence plate in the multi-source data in interval and link interval and the flow in microwave detection data, speed, time occupancy are as the input of Fusion Model, carry out data fusion, obtain the fusion value in the interval and link of licence plate interval.

Optionally, in order to obtain merging more accurately value, the data of described Fusion Model input also comprise: the Floating Car sample size.

The disposal route that the embodiment of the present invention provides a kind of multi-source traffic information to merge is obtained multi-source data, and described multi-source data is carried out pre-service, will described pretreated multi-source data matches respectively the interval and link of licence plate interval; According to the shortage of data situation of described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model; According to the described Fusion Model of choosing, the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.Owing to and link interval to licence plate interval respectively choosing Fusion Model according to the deletion condition of multi-source data and the method for filling a vacancy of utilization, and carry out data fusion according to the Fusion Model of choosing, make the interval and licence plate of link interval can obtain to merge more accurately value, and then the fusion value that makes the licence plate interval can reflect the road conditions in long highway section accurately, traffic information accurately is provided, the fusion value in link interval can reflect the road conditions of shorted segment accurately, provides accurately the path to induce.

Embodiment three,

The treating apparatus that the embodiment of the present invention provides a kind of multi-source traffic information to merge, as shown in Figure 3, described device 30 comprises: acquiring unit 31, pretreatment unit 32, matching unit 33, choose unit 34, integrated unit 35.

Described acquiring unit 31 is used for obtaining multi-source data, and described multi-source data comprises floating car data, microwave data.

Fixed detector on city expressway comprises microwave detector and licence plate detecting device.Also exist mobile Floating Car for detection of traffic data on city expressway.

Wherein, the traffic data precision that the licence plate detecting device obtains is high, but the negligible amounts of licence plate detecting device, the data volume of acquisition is less, and the transport information that can reflect is limited.But for the highway section that there is no the licence plate detecting device, if only utilize single detecting device, detect or the detection of Floating Car detecting device as microwave, the data of detection can accurately not reflect traffic information.Therefore, carry out the fusion of multi-source traffic information, transport information is necessary more accurately in acquisition.

Wherein, the data volume of the floating car data that the Floating Car detection system detects is more, can reflect transport information in a big way.Therefore, in the embodiment of the present invention, adopt the microwave data of floating car data and microwave detector detection as multi-source data, carry out the fusion of multi-source traffic information, obtain wide coverage, and the higher traffic data of precision.

Wherein, the data of Floating Car detection system acquisition comprise: Floating Car Average Travel Speed and Floating Car sample size.Wherein, the Floating Car Average Travel Speed represents the average overall travel speed in corresponding link interval; The Floating Car sample size represents the quantity of the Floating Car of travelling on corresponding link.

The data that microwave detector obtains comprise: flow, speed and time occupancy.Wherein, flow represents in the unit interval by the vehicle number between a certain microwave region; Speed represents the vehicle instantaneous velocity mean value by detecting device; Time occupancy represents time that microwave detector in specific observation time is taken by vehicle and the ratio of observation time.Time occupancy hour, the vehicle by microwave detector in the unit interval is less, and travel speed is higher, causes time occupancy lower; When time occupancy was larger, the vehicle by microwave detector in the unit interval was more, and the speed of travelling is lower, causes time occupancy higher.

Concrete, described acquiring unit 31 can obtain floating car data by the Floating Car detection system, can obtain microwave data by microwave detector.

Described pretreatment unit 32, the multi-source data that is used for described acquiring unit is obtained carries out pre-service, and described pre-service comprises that conversion is unified, misdata is rejected, data are filled a vacancy and time match.

After described acquisition unit 31 obtains described multi-source datas, described pretreatment unit 32 need to be changed unification to described multi-source data, make described multi-source data can be expressed as the data of unified parameter, the data of described unified parameter can be: speed, time occupancy etc. can reflect the parameter of road conditions.The embodiment of the present invention is all unified floating car data, microwave data and licence plate data to characterize with identical parameters such as flow, speed, time occupancies.

Change after reunification, described floating car data comprises: link number, date, timestamp, interval travelling speed.Described microwave data comprises: detecting device numbering, date, timestamp, flow, speed, time occupancy.

Misdata refers to have occured do not meet the traffic parameter data of convention sudden change in certain independent sampling interval, normally because detecting device, transmission line failure and vehicles failed cause by the traffic control reason such as travel.When traffic detecting device or transmission line broke down, the data that collect were normally wrong, can not reflect real traffic.The embodiment of the present invention is rejected after adopting threshold method and traffic flow theory to judge to the processing of misdata.

Adopting threshold method and traffic flow theory that misdata is rejected, is technology well known to those skilled in the art, and the embodiment of the present invention no longer specifically describes at this.

Shortage of data can be divided into two classes: a class is the inherent data disappearance, this is because the detecting device sweep frequency is fixing, equipment breaks down, the excessive intensive detecting device that causes of vehicle can't normally detect the reasons such as vehicle, the dynamic data that collects can't be uploaded in strict accordance with the fixed time interval, the shortage of data that causes; Another kind of is the shortage of data that causes after misdata is rejected.

For the situation of shortage of data, described pretreatment unit 32 is filled a vacancy to missing data according to the method for filling a vacancy.The described method of filling a vacancy comprises: time series method, historical data base method and the spatial position data method of filling a vacancy of filling a vacancy of filling a vacancy.

In prior art, adopt time series the fill a vacancy order of method of method-historical data base of filling a vacancy that the floating car data of disappearance is carried out data and fills a vacancy for floating car data more; Adopt time series method-spatial position data fill a vacancy order of method of method-historical data base of filling a vacancy of filling a vacancy that the microwave data of disappearance is carried out data and fills a vacancy for microwave data more.

In the embodiment of the present invention, in licence plate interval and the interval fusion value more accurately that obtains of link, described pretreatment unit 32 is selected the different methods of filling a vacancy for different shortage of data situations for respectively.

When floating car data lacked, the type of the Linktype corresponding according to missing data and the adjacent link of described link was determined the method for filling a vacancy of the multi-source data of missing data.

When microwave data lacks, the microwave detector corresponding according to missing data respectively with correlativity ρ 1 and the ρ 2 of adjacent two microwave detectors, determine the method for filling a vacancy of the multi-source data of missing data;

Wherein, $\mathrm{\ρ}1,\mathrm{\ρ}2=\frac{\mathrm{cov}(x,y)}{\sqrt{D\left(x\right)}\sqrt{D\left(y\right)}}=\frac{E((x-E\left(x\right))*(y-E\left(y\right)))}{\sqrt{D\left(x\right)}\sqrt{D\left(y\right)}};$

For p1, x, y represent that respectively microwave detector and upstream microwave detector that in historical data base, missing data is corresponding detect the same time sequence traffic information data vector that obtains; For p2, x, y represent that respectively microwave detector and downstream microwave detecting device that in historical data base, missing data is corresponding detect the same time sequence traffic information data vector that obtains.

To the method that floating car data and microwave data are filled a vacancy according to the situation of missing data, referring to the specific descriptions in the embodiment of the present invention two, the embodiment of the present invention does not repeat them here.

Because each checkout equipment works alone, the interval of the data that each checkout equipment detects is also not necessarily identical, and therefore, described pretreatment unit 32 need to carry out time match to described multi-source data.The time match of multi-source data refers to that the multi-source traffic flow data that will collect mates, and makes the transport information of data in the same time period of reflection.

Multi-source data is carried out time match, is technology well known to those skilled in the art, and the embodiment of the present invention is not described specifically at this.

Described matching unit 33, licence plate is interval and link is interval for the pretreated multi-source data of described pretreatment unit is matched respectively.

The licence plate interval is the fundamental space unit of the data of licence plate detecting device detection, is the highway section with directivity that forms after according to certain rule, road being divided.

The link interval is the fundamental space unit of the data of Floating Car detection system detection, is the highway section with directivity that forms after according to certain rule, road being divided.

The fundamental space unit of the data that between the microwave region, microwave detector detects is the highway sections with directivity that form after according to certain rule, road being divided.

Need to prove, generally, the licence plate interval comprises between a plurality of microwave regions, and comprises between each microwave region that a plurality of links are interval.

Therefore the traffic information that can accurately reflect long highway section due to the data in licence plate interval, in the multi-source traffic information integration technology, matches the licence plate interval with the multi-source data that obtains, and makes multi-source data can reflect accurately that the road conditions in long highway section are necessary.

The traffic information that can reflect accurately shorted segment due to the data in link interval, therefore, in the multi-source traffic information integration technology, the multi-source data that obtains is matched the link interval, make multi-source data can reflect accurately the road conditions of shorted segment, can further provide accurately the path to induce.

The method that floating car data and microwave data is matched the licence plate interval comprises: at first, floating car data is converted to link zone-to-zone travel stream parameter, microwave data is converted to traffic flow parameter between the microwave region; Obtain respectively between link interval, microwave region and account for licence plate length of an interval degree ratio; According to formula (2), described floating car data and microwave data are matched the licence plate interval.Wherein, formula (2) is:

$A_{\mathrm{PL}}=\frac{x_1}{L}\×A_1+\frac{x_2}{L}\×A_2+\·\·\·+\frac{x_i}{L}\×A_i;$ Formula (2)

In formula (2), the i indication board is according to interval i the microwave detector interval that comprises or link; Traffic flow parameter after the APL indication board mates according to interval space can be speed, the volume of traffic and occupation rate; The Ai indication board is according to the traffic flow parameter of interval interior i microwave detector interval or link detecting; The xi indication board is according to the length (m) of interval interior i microwave detector interval or link; The L indication board is according to burst length (m).Floating car data and microwave data are matched the link interval.

The floating car data that the Floating Car detection system obtains is take the link interval as unit, therefore, when multi-source data being matched link when interval, do not need floating car data is mated.Microwave data is matched link when interval, owing to comprising between the microwave region that link is interval, therefore, the data of the microwave data between the microwave region as the link interval that comprises between the microwave region are got final product.Link for crossing over two microwave detectors mates according to formula (2).

The described unit 34 of choosing is used for the shortage of data situation according to described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, and the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model.

The described method of filling a vacancy that adopts when fill a vacancy according to the deletion condition of described multi-source data and data in unit 34 of choosing is for the multi-source data in licence plate interval is chosen the interval Fusion Model of licence plate.The described method of filling a vacancy that adopts when fill a vacancy according to the deletion condition of described multi-source data and data in unit 34 of choosing is for the multi-source data in link interval is chosen the interval Fusion Model of link.

Described integrated unit 35 is used for according to the described Fusion Model of choosing unit selection, and the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.

The Fusion Model of choosing according to the licence plate interval, Floating Car Average Travel Speed in the multi-source data in the interval and link of the described licence plates of 35 pairs of described integrated units interval and the flow in microwave detection data, speed, time occupancy are as the input of Fusion Model, carry out data fusion, obtain the fusion value in the interval and link of licence plate interval.

Optionally, in order to obtain merging more accurately value, the data of described Fusion Model input also comprise: the Floating Car sample size.

Before carrying out multisource data fusion, at first need the shortage of data situation that may occur according to multi-source data and the method for filling a vacancy used, obtain many groups of the first multi-source datas of the method for filling a vacancy of every kind of multi-source data deletion condition and correspondence, and according to described many groups the first multi-source data, be deletion condition and the corresponding method of the filling a vacancy training Fusion Model that adopts of every kind of multi-source data.

As shown in Figure 3, described device 30 also comprises: determining unit 36.

Described determining unit 36 is used for according to the deletion condition of the first multi-source data and the method for filling a vacancy accordingly, is respectively interval definite corresponding the first Fusion Model of licence plate interval and link.

Described the first multi-source data comprises: the first floating car data and the first microwave data.The deletion condition of described the first multi-source data is for lacking or not lacking.The described method of filling a vacancy comprises: time series method, historical data base method and the spatial position data method of filling a vacancy of filling a vacancy of filling a vacancy.

For deletion condition and the corresponding method of filling a vacancy of every kind of multi-source data, according to the deletion condition of the first multi-source data and the method for filling a vacancy accordingly, be respectively interval definite corresponding the first Fusion Model of licence plate interval and link.

For instance, when the floating car data in the first multi-source data and microwave data do not lack, be respectively interval and interval definite the first Fusion Model of link of licence plate.

Described acquiring unit 31 also is used to described the first Fusion Model to choose many groups the first multi-source data.

Described acquiring unit 31 is chosen many groups many group first multi-source datas corresponding with described the first Fusion Model.For example, when the floating car data in the first multi-source data corresponding to described the first Fusion Model and microwave data do not lack, the first multi-source data when choosing many group Floating Car and microwave data and fill a vacancy for described the first Fusion Model.

Described pretreatment unit 32 also is used for described many groups the first multi-source data is carried out pre-service.

Described matching unit 33 is used for that also described pretreated many groups the first multi-source data is matched respectively the licence plate interval or link is interval.

Described integrated unit 35 also is used for according to described many groups the first multi-source datas, described the first Fusion Model being carried out model training, obtains a plurality of second Fusion Model in the interval and link of described licence plate interval and the fusion value of correspondence.

Described integrated unit 35 utilizes described many groups the first multi-source data to carry out model training to described the first Fusion Model, can obtain a plurality of second Fusion Model in the interval and link of licence plate interval and the fusion value of correspondence.

Described determining unit 36, also be used for obtaining described licence plate interval or link interval as the average relative error of the licence plate data of true value and the fusion value of least error sum-of-squares difference minimum, second Fusion Model corresponding to fusion value of the described difference minimum in described licence plate interval or link interval is defined as the Fusion Model of the multi-source data in described licence plate interval or link interval.

The average relative error of a plurality of fusion values that the described determining unit 36 described true value of calculating and described a plurality of the second Fusion Model are corresponding, obtain fusion value corresponding to described a plurality of the second Fusion Model and the difference of described true value, determine the precision of described a plurality of the second Fusion Model.

The least error quadratic sum of a plurality of fusion values that the described determining unit 36 described true value of calculating and described a plurality of the second Fusion Model are corresponding, obtain fusion value corresponding to described a plurality of the second Fusion Model and the difference of described true value, determine the validity of described a plurality of the second Fusion Model.

The least error quadratic sum of a plurality of fusion values that the described determining unit 36 described true value of calculating and described a plurality of the second Fusion Model are corresponding, obtain fusion value corresponding to described a plurality of the second Fusion Model and the difference of described true value, determine that the validity of described a plurality of the second Fusion Model specifically comprises: calculate without the first multi-source data of data fusion and the least error quadratic sum between described true value; If the least error quadratic sum between described fusion value and described true value is less than the least error between the floating car data in described the first multi-source data or microwave data and described true value square, described the second Fusion Model is effective.

Described determining unit 36 is determined effective the second model in a plurality of the second Fusion Model, determines that a plurality of second the second the highest Fusion Model of Fusion Model precision are the Fusion Model in described licence plate interval or link interval.

The treating apparatus that the embodiment of the present invention provides a kind of multi-source traffic information to merge, described acquiring unit obtains multi-source data; Described pretreatment unit carries out pre-service to described multi-source data; Described matching unit matches respectively the interval and link of licence plate interval with described pretreated multi-source data; The described unit of choosing is according to the shortage of data situation of described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, and the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model; Described determining unit is according to the described Fusion Model of choosing, and the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.Owing to and link interval to licence plate interval respectively choosing Fusion Model according to the deletion condition of multi-source data and the method for filling a vacancy of utilization, and carry out data fusion according to the Fusion Model of choosing, make the interval and licence plate of link interval can obtain to merge more accurately value, and then the fusion value that makes the licence plate interval can reflect the road conditions in long highway section accurately, traffic information accurately is provided, the fusion value in link interval can reflect the road conditions of shorted segment accurately, provides accurately the path to induce.

One of ordinary skill in the art will appreciate that: all or part of step that realizes said method embodiment can be completed by the hardware that programmed instruction is correlated with, aforesaid program can be stored in a computer read/write memory medium, this program is carried out the step that comprises said method embodiment when carrying out; And aforesaid storage medium comprises: the various media that can be program code stored such as ROM, RAM, magnetic disc or CD.

The above; be only the specific embodiment of the present invention, but protection scope of the present invention is not limited to this, anyly is familiar with those skilled in the art in the technical scope that the present invention discloses; can expect easily changing or replacing, within all should being encompassed in protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of described claim.

Claims (8)

1. the disposal route that multi-source traffic information merges, is characterized in that,

Obtain multi-source data, described multi-source data comprises floating car data, microwave data, licence plate data;

Described multi-source data is carried out pre-service, and described pre-service comprises that conversion is unified, misdata is rejected, data are filled a vacancy and time match;

Described pretreated multi-source data is matched respectively the interval and link of licence plate interval;

According to the shortage of data situation of described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model;

According to the described Fusion Model of choosing, the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.

2. method according to claim 1, is characterized in that, described data fill a vacancy into:

When described floating car data disappearance, the type of the Linktype corresponding according to missing data and the adjacent link of described link is determined the method for filling a vacancy of missing data;

When microwave data lacked, the microwave detector corresponding according to missing data be correlativity p1 and the p2 of two microwave detectors adjacent with upstream and downstream respectively, determined the method for filling a vacancy of missing data;

Wherein, $p1,p2=\frac{\mathrm{cov}(x,y)}{\sqrt{D\left(x\right)}\sqrt{D\left(y\right)}}=\frac{E((x-E\left(x\right))*(y-E\left(y\right)))}{\sqrt{D\left(x\right)}\sqrt{D\left(y\right)}};$

For p1, x, y represent that respectively microwave detector and upstream microwave detector that in historical data base, missing data is corresponding detect the same time sequence traffic information data vector that obtains; For p2, x, y represent that respectively microwave detector and downstream microwave detecting device that in historical data base, missing data is corresponding detect the same time sequence traffic information data vector that obtains;

According to the described method of filling a vacancy, the multi-source data of described missing data being carried out data fills a vacancy.

3. method according to claim 1 and 2, is characterized in that, according to the described Fusion Model of choosing, the multi-source data in and link interval to described licence plate interval carries out data fusion and is specially respectively:

According to the described Fusion Model of choosing, the speed in the multi-source data in and link interval to described licence plate interval, and the sample size of floating car data respectively, the time occupancy of microwave data, the flow of microwave data carry out data fusion.

4. according to claim 1-3 described methods of any one, it is characterized in that, in the shortage of data situation according to described multi-source data, and the method for filling a vacancy of described multi-source data employing, the multi-source data that is respectively the interval and link of described licence plate interval also comprises before choosing Fusion Model:

According to the deletion condition of the first multi-source data and the method for filling a vacancy accordingly, be respectively interval definite corresponding the first Fusion Model of described licence plate interval and link;

For described the first Fusion Model is chosen many groups the first multi-source data;

Described many groups the first multi-source data is carried out pre-service;

Described pretreated many groups the first multi-source data is matched respectively the licence plate interval or link is interval;

According to described many groups the first multi-source datas, described the first Fusion Model is carried out model training, obtain a plurality of second Fusion Model in the interval and link of described licence plate interval and the fusion value of correspondence;

Obtain described licence plate interval or link interval with as the average relative error of the licence plate data of true value and the fusion value of least error sum-of-squares difference minimum, second Fusion Model corresponding to fusion value of the described difference minimum in described licence plate interval or link interval is defined as the Fusion Model of the multi-source data in described licence plate interval or link interval.

5. the treating apparatus that merges of a multi-source traffic information, is characterized in that, described device comprises: acquiring unit, pretreatment unit, matching unit, choose unit, integrated unit;

Described acquiring unit is used for obtaining multi-source data, and described multi-source data comprises floating car data, microwave data;

Described pretreatment unit, the multi-source data that is used for described acquiring unit is obtained carries out pre-service, and described pre-service comprises that conversion is unified, misdata is rejected, data are filled a vacancy and time match;

Described matching unit, licence plate is interval and link is interval for the pretreated multi-source data of described pretreatment unit is matched respectively;

The described unit of choosing is used for the shortage of data situation according to described multi-source data, and the method for filling a vacancy that adopts of described multi-source data, and the multi-source data that is respectively the interval and link of described licence plate interval is chosen Fusion Model;

Described integrated unit is used for according to the described Fusion Model of choosing unit selection, and the multi-source data in and link interval to described licence plate interval carries out data fusion respectively, obtains the fusion value in described licence plate interval and link interval.

6. device according to claim 5, is characterized in that, described pretreatment unit specifically is used for:

When described floating car data disappearance, the type of the Linktype corresponding according to missing data and the adjacent link of described link is determined the method for filling a vacancy of missing data;

When microwave data lacked, the microwave detector corresponding according to missing data be correlativity p1 and the p2 of two microwave detectors adjacent with upstream and downstream respectively, determined the method for filling a vacancy of missing data;

Wherein, $p1,p2=\frac{\mathrm{cov}(x,y)}{\sqrt{D\left(x\right)}\sqrt{D\left(y\right)}}=\frac{E((x-E\left(x\right))*(y-E\left(y\right)))}{\sqrt{D\left(x\right)}\sqrt{D\left(y\right)}};$

For p1, x, y represent that respectively microwave detector and upstream microwave detector that in historical data base, missing data is corresponding detect the same time sequence traffic information data vector that obtains; For p2, x, y represent that respectively microwave detector and downstream microwave detecting device that in historical data base, missing data is corresponding detect the same time sequence traffic information data vector that obtains;

According to the described method of filling a vacancy, the multi-source data of described missing data being carried out data fills a vacancy.

7. according to claim 5 or 6 described devices, is characterized in that, described integrated unit specifically is used for:

According to the described Fusion Model of choosing unit selection, the multi-source data in and link interval to described licence plate interval, and the sample size of floating car data respectively, the time occupancy of microwave data, the flow of microwave data carry out data fusion.

8. according to claim 5-7 described devices of any one, is characterized in that, described device also comprises: determining unit;

Described determining unit is used for according to the deletion condition of the first multi-source data and the method for filling a vacancy accordingly, is respectively interval definite corresponding the first Fusion Model of described licence plate interval and link;

Described acquiring unit also is used to described the first Fusion Model to choose many groups the first multi-source data;

Described pretreatment unit, many groups of the first multi-source datas that also are used for described acquiring unit is obtained carry out pre-service;

Described matching unit is used for that also the pretreated many groups of described pretreatment unit the first multi-source data is matched respectively the licence plate interval or link is interval;

Described integrated unit also is used for according to described many groups the first multi-source datas, described the first Fusion Model being carried out model training, obtains a plurality of second Fusion Model in the interval and link of described licence plate interval and the fusion value of correspondence;

Described determining unit, also be used for obtaining described licence plate interval or link interval with as the average relative error of the licence plate data of true value and the fusion value of least error sum-of-squares difference minimum, second Fusion Model corresponding to fusion value of the described difference minimum in described licence plate interval or link interval is defined as the Fusion Model of the multi-source data in described licence plate interval or link interval.

Images