CN102881162B - Data processing and fusion method for large-scale traffic information - Google Patents

Abstract

The invention discloses a large-scale traffic information data processing and fusion method, which belongs to the traffic information real-time processing technology, including: obtaining the standard data of the true value system according to the multi-source traffic data collected by the test vehicle and various sensors, and determining the parameters The dynamic assignment method; eliminate the abnormal data in the data collection collected by the sensor, and compensate the historical data; the real-time classification information fusion of the multi-source traffic data that has completed the data compensation. The present invention obtains the correct initial assignment parameters of various acquisition methods by establishing a truth system, and removes abnormalities from the actually collected data, and fills in missing data reasonably according to historical data to ensure the accuracy and integrity of the data. The data obtained by the collection method is fused step by step to ensure the reliability of the data and the speed of the fusion process, and the impact of traffic events, traffic control, land occupation construction, and traffic accidents on the data is considered during the fusion process.

Description

Data processing and fusion method of large-scale traffic information

technical field

The invention relates to the technical field of real-time processing of traffic information of an intelligent traffic system, in particular to a data processing and fusion method of large-scale traffic information.

Background technique

Information fusion was first applied in the military field. It is a multi-level and multi-faceted process of combining multi-source information and data to complete target detection, association, and status assessment. The purpose of this information fusion is to obtain accurate target identification, complete and timely battlefield situation and threat assessment. With the development of sensor technology, computer science and information technology, a large number of multi-sensor systems for complex application backgrounds have emerged, making multi-channel information acquisition, processing and fusion possible. In many fields such as diagnosis, weather forecast, organization and management decision-making, robot vision, traffic control, remote sensing and telemetry, people have realized that combining information from multiple data sources can improve work performance. Therefore, multi-source information fusion technology has been widely valued and successfully applied in military and civilian fields, and its theory and method have become an important research direction of intelligent information processing and control.

Existing technologies are more focused on data-level information fusion, which directly synthesizes and analyzes raw sensor observation data without preprocessing. The advantage is that it keeps as much object information as possible, and there is basically no information loss or omission; the disadvantage is that it processes too much data, takes too long, and has poor real-time performance. Another type of information fusion is feature-level information fusion, also known as file-level information fusion. It is a comprehensive analysis and processing performed on the basis of the basic feature extraction of the sensor and the provision of file reports after the initial preprocessing of the sensor. Its advantage is that it not only maintains a sufficient amount of important information, but also has undergone allowable data compression, which greatly dilutes the amount of data and can improve the real-time performance of the processing process; and it is particularly valuable in pattern recognition, image analysis, computer vision, etc. In modern high-tech applications, feature extraction is actually the basis, and a lot of work has been carried out in this area. The disadvantage of feature-level information fusion is that there will inevitably be some information loss, so it is necessary to put forward stricter requirements for sensor preprocessing.

Due to the multi-source nature of traffic data, an effective method must be used to coordinate multi-source data reasonably, fully integrate useful information, and improve the ability to make correct decisions in a changing environment. Data fusion technology is a good tool to meet this requirement. The multi-source and complexity of traffic data requires that traffic information processing must have data fusion technology, so as to provide the robustness of traffic information.

The existing traffic data fusion technology is more focused on data-level information fusion. The fusion process does not classify the data according to the stability of the sensor, and does not check and eliminate the data collected by the sensor. The data is fused step by step. The impact of traffic incidents, traffic control, land occupation construction, and traffic accidents on the data is not considered.

Contents of the invention

In view of this, the object of the present invention is to provide a large-scale traffic information data processing and fusion method for realizing efficient fusion of traffic information data.

The present invention provides a data processing and fusion method of large-scale traffic information, including:

Obtain the standard data of the truth system according to the multi-source traffic data collected by the test vehicle and each sensor, and determine the dynamic assignment method of the parameters; eliminate the abnormal data in the data set collected by the sensors, and perform historical data compensation; Real-time hierarchical information fusion of multi-source traffic data that has completed data compensation;

The step of obtaining the standard data of the truth system specifically includes:

Through multiple test vehicles with high sampling rate, multiple tests in different periods, and the average value of period t is obtained as the standard value:

$\mathrm{<span\; class="notranslate">\; Et</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; no</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\frac{\mathrm{<span\; class="notranslate">\; Xij</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; m</span>}}$

Among them, Xi, j(t) represents the measurement value of the i-th test vehicle in the j-th test in the time period t, and n vehicles perform m measurements;

The steps of the dynamic assignment method for determining parameters specifically include:

Split the data collected by the test vehicle and various sensors in time period t into taxi data, bus data, microwave/coil data, video surveillance data and license plate recognition data as various traffic information data in time period t ;

The standard deviation is calculated by taking the average value of taxi data, bus data, microwave/coil data, video surveillance data and license plate recognition data in time period t as the standard value:

$\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\sqrt{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; Xi</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

Among them, E(X) is the average value of all data obtained in the t-period test, and Xi represents the i-th test data of various types of data;

Using the standard value and standard deviation, obtain the normal distribution diagram of each type of data respectively, determine the credible interval, and calculate the probability of the data in the credible interval according to the normal curve:

$\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; \&Integral;</span>}_{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; b</span>}}\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&Pi;</span>}}}{\mathrm{<span\; class="notranslate">\; e</span>}}^{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

Among them, a and b respectively represent the limit value of acceptable data around which the test data deviates from the standard value X, D(X) is the standard deviation, and E(X) is the average value;

The basic probability assignment function m(A) of each type of data is assigned with the probability of the data in the credible interval, the assignment process is m(A)=P(a<A<b), A represents each type of data, P( a<A<b) represents the probability of each type of data in the credible interval.

The present invention obtains the correct initial assignment parameters of various acquisition methods by establishing a truth system, and removes abnormalities from the actually collected data, and fills in missing data reasonably according to historical data to ensure the accuracy and integrity of the data. The data obtained by the collection method is fused step by step to ensure the reliability of the data and the speed of the fusion process, and the impact of traffic events, traffic control, land occupation construction, and traffic accidents on the data is considered during the fusion process. Through this fusion process, the fast, accurate, and reliable effects of fusion are realized. The truth system solves the problem of inaccurate initial parameter assignment in the prior art; the removal of abnormal data and filling of missing data overcomes the unstable performance of the acquisition terminal in the prior art , and the problem of damage to the collection terminal; the D-S theory is used to perform hierarchical fusion processing on the processed data, which solves the multi-source problem of data fusion; also fully considers traffic incidents, traffic control, land occupation construction, traffic accidents, and overcomes the The impact of the above factors on the data.

Description of drawings

FIG. 1 is a flow chart of a data processing and fusion method for large-scale traffic information provided by an embodiment of the present invention;

Fig. 2 is a flowchart of a method for dynamically assigning values to parameters in an embodiment of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings.

The traffic data fusion process provided by the embodiment of the present invention can achieve fast, accurate, reliable and other effects of fusion. The truth system solves the problem of inaccurate initial parameter assignment in the prior art; it overcomes the existing problem of eliminating abnormal data and filling missing data. The performance of technical acquisition terminals is unstable, and the problem of damage to acquisition terminals; the D-S theory is used to perform hierarchical fusion processing on the processed data, which solves the multi-source problem of data fusion; it also fully considers traffic incidents, traffic control, land occupation construction, Traffic accidents overcome the influence of the above factors on the data.

Fig. 1 is a flow chart of a data processing and fusion method for large-scale traffic information provided by an embodiment of the present invention, including the following steps:

Step 101, obtain the truth system, and determine the dynamic assignment method of parameters in D-S (Dempster-Sharer evidence theory multi-source data fusion method).

The standard data of the true value system is tested multiple times at different time periods by multiple test vehicles with high sampling rates, and the average value of the time period t is obtained as the standard value:

$\mathrm{<span\; class="notranslate">\; Et</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; no</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\frac{\mathrm{<span\; class="notranslate">\; Xij</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; m</span>}}$

Xi _,j (t) represents the measurement value of the jth test of the i-th test vehicle at the time period t, and n vehicles make m measurements.

Figure 2 is a flowchart of a method for dynamically assigning values to parameters, including the following steps:

Step 1011, split the data collected by the test vehicle and the data collected by each sensor in time period t into taxi data, bus data, microwave/coil data, video surveillance data, and license plate recognition data, as various types of time period t traffic information data. Since different collection devices can collect the same type of data, but the collection accuracy of each collection device and its influencing factors are different, therefore, there will be step-by-step fusion in the subsequent steps.

Step 1012, calculate the standard deviation using the average value of taxi data, bus data, microwave/coil data, video surveillance data and license plate recognition data in time period t as the standard value:

$\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\sqrt{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; Xi</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

Among them, E(X) is the average value of all data obtained in the t-period test, and _Xi represents the i-th test data of various types of data.

The standard deviation reflects the degree of dispersion of the data, that is, the degree of stability of the data.

Step 1013, use the standard deviation and standard value to obtain the normal distribution map of each type of data respectively, determine the credible interval (that is, generate the data receiving interval), and calculate the probability of the data in the credible interval according to the normal curve:

$\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; \&Integral;</span>}_{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; b</span>}}\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&Pi;</span>}}}{\mathrm{<span\; class="notranslate">\; e</span>}}^{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

Among them, a and b respectively represent the limit value of the acceptable data that the test data deviates from the standard value X, D(X) is the standard deviation, and E(X) is the average value.

Step 1014, assign a value to the basic probability assignment function m(A) of each type of data with the probability that the data is in the credible interval. For example, the basic probability assignment function m(A) of taxi data has an assignment process of m(A)=P(a<A<b), where A represents the taxi data, and P(a<A<b) represents the data collected by the taxi in the available The probability of the confidence interval, calculated as above.

Step 102, eliminating abnormal data in the data set collected by the sensor. Step 101 is the data obtained from multiple tests of the test vehicle under normal conditions, and prepares for the initial parameter assignment of various data in the data fusion process. In the real situation, various factors will lead to abnormal data. Therefore, after the actual sensor collects the data, the abnormal data must be eliminated first. The following is the processing of the actual data.

Calculate the time period t corresponding to the classification (including taxi data, bus data, microwave/coil data, video surveillance data and license plate recognition data) traffic information data X (X is the data collected in the actual environment, and step 101 describes X in The data under the normal environment of the test vehicle and the collector) and the historical information data H. H is the storage of traffic information data (including taxi data, bus data, microwave/coil data, video surveillance data and license plate recognition data) in the past year. Calculate the mathematical expectations E(X) and E(H) of X and H, and calculate the standard deviation D(X) and D(H) of various traffic information data X and the historical information data H of the same period in the past year during the calculation period t, if:

(1) For the traffic information data collected by each sensor and test vehicle in the time period t, the difference between the mathematical expectation E(X) and the mathematical expectation E(H) is within the set threshold th, and D(X) and D( H) If the difference is also within the set threshold th', it is considered that there is no abnormality in the data at this moment;

(2) For the traffic information data collected by each sensor and test vehicle in time period t, the mathematical expectation E(X) and mathematical expectation E(H) exceed the set threshold th, but the difference between D(X) and D(H) Within the set threshold th', wait for the traffic information data X" of the next period to be generated; if there is a large difference between the traffic information data D(X) and D(X") of the next period, it exceeds the set threshold th′, it is considered that the traffic information of time period t is abnormal, and the abnormal data will be removed.

In a preferred embodiment, the th threshold is set as th=E(H)/2, th'=D(X)/2.

Step 103, compensation of historical data. The traffic information data collected by each sensor and the test vehicle every day is stored in step 102 as historical data, and used as compensation data in case of data loss in the future.

According to the historical data, calculate the average speed of the road network of various traffic information data in the same day, including the average speed of the road network of taxi data, the average speed of bus data, and the average speed of microwave/coil data, and generate a traffic trend curve.

After each sensor and test vehicle collects data every day, during the process of preliminary data processing (removing abnormal data in the data) in time, search for Q traffic data files contained in adjacent Z time periods (that is, the previously classified storage) Traffic data with similar traffic trend curves), calculate the average speed of the road network of the traffic data file in each time period, and regard the average speed as a point, and smoothly connect Z points to generate a traffic information curve segment. Match the curve segment with the historical data in the same time period, and find the most similar P curves as auxiliary information sources, where Z, P and Q are obtained based on algorithms, historical statistics, practice, experience, and system performance Function value, Z represents the time period, Q represents the number of files generated in Z time period, and P represents the number of most similar curves. Adjacent Z time periods, in units of one hour, 24 time periods. General data collection generates one data every five minutes, 12 data files will be generated in one time period, Q will generate 288 values based on 24 time periods, and the value of P will be 1.

According to the time attributes of various traffic information data collected by each sensor and test vehicle, from the auxiliary information sources obtained in the above steps, find the historical data pair with the same unique number and the same time attributes as the road link with missing traffic data. Compensation area Fill in the gaps in the information link. The time attribute includes the period of the week and the time window characteristic of the moment. The reason why the data will be missing is that on the one hand, the elimination of abnormal data may cause data discontinuity, and on the other hand, due to the short-term failure of each data source device due to external factors or human factors, the data cannot be collected in time or the collection error is large. And the latter is especially important.

Step 104, the multi-source traffic data (the multi-source data here includes taxi data, bus data, microwave/coil data, video surveillance data, and license plate recognition data after elimination and data compensation) is completed in real time Hierarchical information fusion. The first three steps are the basis of real-time classification information fusion. Step 101 generates the true value system for the classification information fusion to determine the fusion parameters of each type of data fusion, and provides standard data for the removal of abnormal data in step 102. Step 103 is for missing data. Data compensation is performed for road links or data missing time periods, and the data generated in steps 102 and 103 are used as the data source of multi-source traffic information data. The data used in the following steps are all processed data in steps 101, 102, and 103.

In order to ensure data accuracy and coverage, large-scale and diverse taxi data and bus data are first fused, and then microwave/coil data and license plate recognition data are fused. Proceed as follows:

The most basic concept in D-S evidence theory is to establish an identification frame, which is denoted as U, and the elements in the identification frame meet the mutually incompatible conditions. The assignment m(A) of proposition A to function m is the mapping to set, if function m satisfies the following conditions:

m(Ф)＝0

right $\&ForAll;A\&Element;2^u,m\left(A\right)\&Greater\; Equal;0,$ and $\underset{A\&Element;2^u}{\mathrm{\&Sigma;}}m\left(A\right)=1$

Then m(A) is called the basic probability assignment function (BPAF) of A. Among them, Ф represents the empty set, and m(A) represents the precise degree of belief in proposition A, expressing the direct support for A.

In the following steps, for the initialization of taxi data, bus data, microwave/coil data, video surveillance data and license plate recognition data, the method of step 1014 is also used to assign values to the basic probability assignment function. Step 1013 obtains the probability of all kinds of data in the credible interval obtained by the test vehicle under normal circumstances, less consideration is given to the impact of emergencies, road maintenance, etc. on the range of collected data values, so the following fusion In the process, emergencies, road maintenance, etc. should also be considered.

Step 1041, firstly integrate taxi and bus data. Due to the high usage rate of devices such as GPS, data accuracy and coverage can be guaranteed.

The initial value of the basic probability assignment function obtained according to the taxi data obtained in step 1014 and the bus data is expressed as m ₁ (A) and m ₂ (A);

Note that the fused data is U, and the taxi data and bus data are _U1 and _U2 respectively;

$\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; *</span>\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; *</span>\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}$

U is the value obtained after the fusion of taxi and bus information, and the precise trust degree of U after fusion is:

$\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\underset{{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; \&cap;</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; A</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}{\underset{{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; \&cap;</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&NotEqual;</span>\mathrm{<span\; class="notranslate">\; \&Phi;</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CirclePlus;</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>$

Step 1042, integrating microwave/coil data. The microwave/coil data further guarantees the stability and reliability of the data.

The initial value m ₃ (A) of the basic probability assignment function of the microwave/coil data obtained according to step 1014;

Note that the fusion data is K, the fusion data obtained in step 1041 is K ₁ (i.e. the U value obtained in 1014), and the microwave/coil data is K ₂ ;

$\mathrm{<span\; class="notranslate">\; K</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; *</span>\frac{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; *</span>\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}$

K is the value obtained after the fusion of taxi, bus and microwave/coil data. The precise trust degree of K after fusion is:

${\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; K</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&cap;</span>\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}{\underset{\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&cap;</span>\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; \&NotEqual;</span>\mathrm{<span\; class="notranslate">\; \&Phi;</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CirclePlus;</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>$

Step 1043, integrating the license plate recognition data. High-definition cameras are generally installed in the intersection area, and the integration of high-definition camera license plate recognition data can ensure the accuracy and reliability of the data passing through the checkpoint road section. The method is the same as step 1042.

Step 1044, integrating the signal light timing data of trunk roads, expressways, and secondary trunk roads. High-definition cameras are generally installed in the intersection area. Due to the collection of vehicles at the intersection, it has a great impact on the speed of the vehicle. The integration of high-definition camera license plate recognition data can ensure the accuracy and reliability of the data passing through the bayonet road section. Trunk roads, expressways, The data between sub-trunk road sections is greatly affected by the timing data of signal lights, and the data accuracy can be further improved by integrating high-definition camera data and signal light timing data respectively. Due to the interference of signal lights, the driving time and speed of vehicles on the road link will change, and the range of collected data values should be appropriately enlarged, that is, the range of interval [a, b] should be appropriately relaxed when calculating in step 1013. The fusion method Same as step 1042.

Step 1045, integrating traffic incidents, traffic control, station road construction, and traffic accident data. Due to the impact of traffic events, traffic control, station road construction, and traffic accidents on the data, the range of data values collected by taxis, buses, microwaves/coils, etc. should be appropriately enlarged, that is, the interval should be appropriately relaxed when calculating in step 1013 [ a, b] range, ibid.

The above methods have fully considered the stability, accuracy and reliability of various acquisition methods. Establish a truth system to obtain the correct initial assignment parameters for various collection methods, remove abnormalities from the real collected data, and fill in missing data reasonably according to historical data to ensure the accuracy and integrity of the data. The data is fused step by step to ensure the reliability of the data and the rapidity of the fusion process, and the impact of traffic incidents, traffic control, land occupation construction, and traffic accidents on the data is considered during the fusion process.

Through the fusion process of the above steps, the fast, accurate and reliable effects of fusion are realized. The truth system solves the problem of inaccurate initial parameter assignment in the prior art; the removal of abnormal data and filling of missing data overcomes the performance of the acquisition terminal in the prior art Instability, and the problem of damage to the collection terminal; the D-S theory is used to perform hierarchical fusion processing on the processed data, which solves the multi-source problem of data fusion. Fully consider traffic incidents, traffic control, land occupation construction, and traffic accidents, and overcome the influence of the above factors on the data.

In a word, the above descriptions are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention.

Claims (7)

1. A data processing and fusion method for large-scale traffic information, characterized in that, comprising: Obtain the standard data of the truth system according to the multi-source traffic data collected by the test vehicle and each sensor, and determine the dynamic assignment method of the parameters; eliminate the abnormal data in the data set collected by the sensors, and perform historical data compensation; Real-time hierarchical information fusion of multi-source traffic data that has completed data compensation; The step of obtaining the standard data of the truth system specifically includes: Through multiple test vehicles with high sampling rate, multiple tests in different periods, and the average value of period t is obtained as the standard value: $\mathrm{<span\; class="notranslate">\; Et</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; no</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\frac{\mathrm{<span\; class="notranslate">\; Xij</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; m</span>}}$ Among them, X _{i, j} (t) represents the measured value of the jth test of the i-th test vehicle in the time period t, and n vehicles perform m measurements; The steps of the dynamic assignment method for determining parameters specifically include: Split the data collected by the test vehicle and various sensors in time period t into taxi data, bus data, microwave/coil data, video surveillance data and license plate recognition data as various traffic information data in time period t ; The standard deviation is calculated by taking the average value of taxi data, bus data, microwave/coil data, video surveillance data and license plate recognition data in time period t as the standard value: $\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\sqrt{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; Xi</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$ Among them, E(X) is the average value of all data obtained in the t-period test, and _Xi represents the i-th test data of various types of data; Using the standard value and standard deviation, obtain the normal distribution diagram of each type of data respectively, determine the credible interval, and calculate the probability of the data in the credible interval according to the normal curve: $\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; \&Integral;</span>}_{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; b</span>}}\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&Pi;</span>}}}{\mathrm{<span\; class="notranslate">\; e</span>}}^{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$ Among them, a and b respectively represent the limit value of acceptable data around which the test data deviates from the standard value X, D(X) is the standard deviation, and E(X) is the average value; The basic probability assignment function m(A) of each type of data is assigned with the probability of the data in the credible interval, the assignment process is m(A)=P(a<A<b), A represents each type of data, P( a<A<b) represents the probability of each type of data in the credible interval. 2. The data processing and fusion method of large-scale traffic information according to claim 1, wherein the step of removing the abnormal data in the data set collected by the sensor specifically comprises: Calculate time period t corresponding to classified traffic information data X and historical information data H, calculate mathematical expectations E(X) and E(H) of X and H, calculate time period t of various traffic information data X and historical information data of the same period in the past year The standard deviation D(X), D(H) of H, if: The difference between mathematical expectation E(X) and mathematical expectation E(H) is within the set threshold th, and the difference between D(X) and D(H) is also within the set threshold th', then the data at this time no exception occurred; Mathematical expectation E(X) and mathematical expectation E(H) exceed the set threshold value th, but the difference between D(X) and D(H) is within the set threshold value th′, then wait for the traffic information data of the next period X" is generated; if the difference between the traffic information data D(X) and D(X") of the next time period exceeds the set threshold th', the traffic information of time period t will be abnormal, and the abnormal data will be removed. 3. The data processing and fusion method of large-scale traffic information according to claim 2, characterized in that, the threshold th and th' are set as: th=E(H)/2, th'=D( X)/2. 4. The data processing and fusion method of large-scale traffic information according to claim 2 or 3, wherein the compensation step of the historical data specifically comprises: After each sensor and test vehicle collects data every day, search for Q traffic data files contained in adjacent Z time periods, calculate the average speed of the road network for each time period traffic data file, and regard the average speed as a point, connect Z points smoothly to generate a traffic information curve segment; match the curve segment with historical data in the same time period, and find the most similar P curves as auxiliary information sources, where Z represents the time period and Q represents The number of files generated in Z time periods, P represents the number of most similar curves; According to the time attributes of various traffic information data collected by each sensor and test vehicle, find historical data with the same unique number and the same time attributes as the link with the lack of traffic data from the auxiliary information source to compensate for the vacancy information in the area Links are filled. 5. The data processing and fusion method of large-scale traffic information according to claim 4, characterized in that, the step of merging real-time graded information of multi-source traffic data that completes data compensation specifically includes fusing taxi and bus data : Obtain the initial value of the basic probability assignment function of taxi data and bus data, denoted as m ₁ (A) and m ₂ (A); Suppose the value obtained after the fusion of taxi and bus data is U, and the taxi data and bus data are U1 and U2 respectively; then $\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; *</span>\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; *</span>\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}$ The precise trust degree of U is: $\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\underset{{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; \&cap;</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; A</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}{\underset{{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; \&cap;</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&NotEqual;</span>\mathrm{<span\; class="notranslate">\; \&Phi;</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CirclePlus;</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span>$ 6. The data processing and fusion method of large-scale traffic information according to claim 5, characterized in that, the step of the real-time hierarchical information fusion of multi-source traffic data that completes data compensation further includes fusing microwave/coil data: Obtain the initial value of the basic probability assignment function for the microwave/coil data, denoted as m ₃ (A); Set the fused data to be K, the value obtained after the fusion of the taxi and bus data is K ₁ , and the microwave/coil data is K ₂ ; $\mathrm{<span\; class="notranslate">\; K</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; *</span>\frac{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; *</span>\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; )</span>}$ K is the value obtained after the fusion of taxi, bus and microwave/coil data. The precise trust degree of K after fusion is: ${\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; K</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&cap;</span>\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}{\underset{\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&cap;</span>\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; \&NotEqual;</span>\mathrm{<span\; class="notranslate">\; \&Phi;</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CirclePlus;</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span>$ 7. The data processing and fusion method of large-scale traffic information according to claim 6, characterized in that, the step of merging real-time graded information of multi-source traffic data for data compensation further includes fusing license plate recognition data, video surveillance Data and traffic incidents, traffic control, station road construction, traffic accident data.