A vehicle code fitting method and system for highway scenes

Abstract

The invention discloses a vehicle code fitting method and system for highway scenarios, comprising: obtaining highway vehicle trajectory data and mobile phone positioning data, extracting key fields, and respectively creating a vehicle spatiotemporal trajectory table and a mobile phone number spatiotemporal trajectory table; performing vehicle code fitting calculations according to the two types of spatiotemporal trajectory tables to obtain daily vehicle code fitting results; further, based on the daily vehicle code fitting results, using threshold filtering and head and tail trajectory-based sample filtering two types of noise filtering methods to reduce the noise of the results to obtain a high-confidence accompanying result candidate set; based on the high-confidence accompanying result candidate set, aggregating monthly to obtain the final vehicle code fitting result; the invention can not only effectively process the noise in the data, but also accurately identify the accompanying relationship between highway vehicles and mobile phone numbers in a sparse data environment, thereby improving the accuracy and credibility of the accompanying relationship identification.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and more particularly to a vehicle code fitting method and system for highway scenarios. Background Art

[0002] In modern traffic management and monitoring systems, vehicle trajectory data and mobile phone positioning data not only help monitor the vehicle's driving trajectory and location, but can also be used for traffic flow analysis, accident detection, illegal behavior monitoring, etc. However, massive vehicle trajectory data and mobile phone positioning data are often sparse and incomplete in time and space, which brings huge challenges to data management, analysis and application.

[0003] Especially in highway scenarios, due to the high-speed movement of vehicles and the complexity of the road environment, highway vehicle trajectory data and mobile phone positioning data are more prone to noise. For example, building occlusion, signal reflection and multipath effects will cause the accuracy of mobile phone positioning data to decrease. In addition, due to the dynamic changes in vehicle density on highways, the complexity of collision detection in trajectory data increases, and traditional longitude-based accompanying relationship recognition methods are difficult to accurately and effectively identify the accompanying relationship between vehicles.

[0004] In the prior art, most of the accompanying relationship recognition methods rely on fixed time and space windows and determine the accompanying relationship through simple trajectory collision detection. However, these methods perform poorly in high-density and high-noise environments, and the accuracy and reliability of the recognition results are difficult to guarantee.

[0005] Therefore, how to effectively process the noise of vehicle trajectory data and mobile phone positioning data, and accurately identify the accompanying relationship in a sparse data environment to improve the accuracy and credibility of accompanying relationship identification, is a problem that technical personnel in this field urgently need to solve. Summary of the invention

[0006] In view of this, the present invention provides a vehicle code fitting method and system for highway scenarios to solve some of the technical problems mentioned in the background technology.

[0007] In order to achieve the above object, the present invention adopts the following technical solution:

[0008] A vehicle code fitting method for a highway scene comprises the following steps:

[0009] S1. Obtain highway vehicle trajectory data and mobile phone positioning data, extract key fields, and create a vehicle spatiotemporal trajectory table and a mobile phone number spatiotemporal trajectory table respectively;

[0010] S2. Car code fitting calculation is performed according to the vehicle spatiotemporal trajectory table and the mobile phone number spatiotemporal trajectory table to obtain the daily car code fitting results;

[0011] S3. Based on the daily vehicle code fitting results, two types of noise filtering methods, threshold filtering and sample filtering based on the first and last trajectories, are used to reduce the noise of the results to obtain a high-confidence accompanying result candidate set;

[0012] S4. Based on the high-confidence accompanying result candidate set, the final vehicle code fitting result is obtained by monthly aggregation.

[0013] Preferably, the key fields of the highway vehicle trajectory data include the vehicle unique identifier ID, the longitude of the highway checkpoint, the latitude of the highway checkpoint, the 12-digit geographic hash value of the highway checkpoint, and the timestamp and date of passing the current checkpoint.

[0014] Preferably, the key fields of the mobile phone positioning data include the unique identifier UID of the person to whom it belongs, the longitude of the mobile phone positioning, the latitude of the mobile phone positioning, the 12-digit geographic hash value of the current location of the mobile phone number, and the timestamp and date of the current location of the mobile phone number.

[0015] Preferably, the specific contents of the vehicle code fitting calculation include:

[0016] S21. Based on the Haversine formula, the distance between the spatiotemporal trajectory point and the associated number trajectory point of each vehicle under each 7-bit timestamp and 5-bit geo-hash value is calculated day by day;

[0017] S22. Find the closest mobile phone number trajectory point and the closest distance to each time and space trajectory point of each vehicle day by day;

[0018] S23. The nearest distance is divided into bins according to the vehicle, mobile phone number, 5-digit geographic hash value of the vehicle, and 7-digit timestamp, and different bins correspond to different scores;

[0019] S24. Aggregate the accompanying scores of all time and space points of each vehicle every month on a daily basis, and retain the top N accompanying mobile phone number scores with the highest scores on a single day, which is the daily vehicle code fitting calculation result.

[0020] Preferably, the specific content of the threshold filtering is:

[0021] Based on the vehicle code fitting calculation results, the scores of the first N accompanying mobile phone numbers of each vehicle on a daily basis are converted into scores based on the z-score, and all mobile phone numbers with a z-score score lower than 1 are filtered out;

[0022] Based on the z-score, the scores are normalized and first-order differences are performed. The point with the largest difference value is taken as the cliff point. All results after the cliff point are filtered out, and all results before the cliff are retained as the final accompanying results.

[0023] Preferably, the specific content of the sample filtering based on the head and tail trajectories is:

[0024] Based on the vehicle code accompanying candidate set obtained by threshold filtering, mark all trajectories with the timestamps of the vehicle code accompanying matching pairs with preset proportion values ​​at the beginning and end, and mark the beginning and end trajectories as 0 and 1 respectively, with the beginning as 0 and the end as 1;

[0025] The vehicle and number are associated based on their 7-digit geo-hash timestamps, and all numbers that only match 0 or only match 1 are filtered out. The mobile phone numbers that match both 0 and 1 are retained, which are the high-level signal code sets that accompany the vehicle code.

[0026] Preferably, the specific content of step S4 is: aggregating the daily accompanying results of each vehicle above the threshold number of days on a monthly basis, and using the candidate number set as the mobile phone number for vehicle code fitting.

[0027] A vehicle code fitting system for highway scenes, based on the vehicle code fitting method for highway scenes, comprises: a spatiotemporal trajectory table creation module, a vehicle code fitting calculation module, a noise filtering module and a result output module;

[0028] The spatiotemporal trajectory table creation module is used to obtain highway vehicle trajectory data and mobile phone positioning data, extract key fields, and create vehicle spatiotemporal trajectory tables and mobile phone number spatiotemporal trajectory tables respectively;

[0029] The vehicle code fitting calculation module is used to perform vehicle code fitting calculation according to the vehicle time-space trajectory table and the mobile phone number time-space trajectory table to obtain the daily vehicle code fitting results;

[0030] The noise filtering module is used to reduce the noise of the daily companion fitting results by using two types of noise filtering methods: threshold filtering and sample filtering based on the head and tail trajectories, to obtain a high-confidence companion result candidate set;

[0031] The result output module is used to aggregate the high-confidence accompanying result candidate set by month to obtain the final vehicle code fitting result.

[0032] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a vehicle code fitting method for a highway scenario.

[0033] A processing terminal includes a memory and a processor. The memory stores a computer program that can be run on the processor. When the processor executes the computer program, a vehicle code fitting method for a highway scene is implemented.

[0034] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a vehicle code fitting method and system for highway scenes, which uses the bayonet trajectory data formed by the highway vehicle bayonet and the mobile phone positioning data as data sources, performs fusion calculation, and obtains the accompanying number of the corresponding vehicle, that is, "finding the code by the vehicle";

[0035] The present invention uses a long-term companion algorithm to perform association calculations on a monthly basis, calculates high-confidence vehicle code fitting results on a daily basis, and obtains accompanying mobile phone number fitting results aggregated on a monthly basis on this basis;

[0036] The present invention can not only effectively process the noise in the data, but also accurately identify the long-term accompanying relationship between vehicles and numbers in a sparse data environment, thereby improving the accuracy and credibility of the accompanying relationship identification and providing strong support for traffic management and monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0038] Figure 1 A schematic diagram of a vehicle code fitting method for a highway scene provided by the present invention;

[0039] Figure 2 A schematic diagram of a vehicle code fitting system for a highway scenario provided by the present invention. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] The embodiment of the present invention discloses a vehicle code fitting method for a highway scene, such as Figure 1 , including the following steps:

[0042] S1. Obtain highway vehicle trajectory data and mobile phone positioning data, extract key fields, and create a vehicle spatiotemporal trajectory table and a mobile phone number spatiotemporal trajectory table respectively;

[0043] S2. Car code fitting calculation is performed according to the vehicle spatiotemporal trajectory table and the mobile phone number spatiotemporal trajectory table to obtain the daily car code fitting results;

[0044] S3. Based on the daily companion fitting results, two types of noise filtering methods, threshold filtering and sample filtering based on the head and tail trajectories, are used to reduce the noise of the results to obtain a high-confidence companion result candidate set;

[0045] S4. Based on the high-confidence accompanying result candidate set, the final vehicle code fitting result is obtained by monthly aggregation.

[0046] In this embodiment, after obtaining the highway vehicle trajectory data and mobile phone positioning data, data extraction, cleaning, and completion operations are performed in sequence to ensure data quality and consistency.

[0047] In order to further implement the above technical solution, the key fields of highway vehicle trajectory data include the vehicle unique identifier ID, the longitude of the highway checkpoint, the latitude of the highway checkpoint, the 12-digit geographic hash value of the highway checkpoint, and the timestamp and date of passing the current checkpoint, where the 12-digit geographic hash value is calculated through the longitude and latitude.

[0048] In order to further implement the above technical solution, the key fields of mobile phone positioning data include the unique identifier UID of the person, the longitude of the mobile phone number track point, the latitude of the mobile phone number track point, the 12-digit geographic hash value of the mobile phone number track point, and the timestamp and date of the mobile phone number track point, where the 12-digit geographic hash value is calculated through longitude and latitude.

[0049] In order to further implement the above technical solution, the specific contents of the vehicle code fitting calculation include:

[0050] S21. Based on the Haversine formula, the distance between the spatiotemporal trajectory point and the associated number trajectory point of each vehicle under each 7-bit timestamp and 5-bit geo-hash value is calculated day by day;

[0051] S22. Find the closest mobile phone number trajectory point and the closest distance to each time and space trajectory point of each vehicle day by day;

[0052] S23. The nearest distance is divided into bins according to the vehicle, the accompanying mobile phone number, the vehicle's 5-digit geographic hash value, and the 7-digit timestamp, and different bins correspond to different scores;

[0053] S24. Aggregate the final scores of each vehicle at all time and space points on each day of a month on a daily basis, and retain the highest scores of the top N accompanying mobile phone numbers on a single day, which is the daily vehicle code fitting calculation result.

[0054] Specific:

[0055] In the highway vehicle trajectory table and mobile phone number spatiotemporal trajectory table, the 7-digit timestamp is intercepted and recorded as time7, and the 5-digit geographic hash value is intercepted and recorded as geo5;

[0056] Associate the vehicle trajectory table time7 and geo5 with the number trajectory table time7 and geo5, calculate the Haversine sphere distance between each spatial trajectory point of each vehicle and each trajectory point of the associated number, and retain the number trajectory point closest to the vehicle's spatiotemporal trajectory point;

[0057] The distances are binned according to the intervals {(0,10],(10,100],(100,200],(200,300],(300,400],(400,500],(500,1000],(1000,2000],(2000,5000],(5000,+∞)}. The scores after binning are [9,8,7,6,5,4,3,2,1,0], and the binning field is recorded as matchscore.

[0058] Finally, the final match scores of each vehicle under geo5 and time7 are aggregated daily, and the top 10 cumulative bin scores per month are counted, which is the vehicle code fitting calculation result.

[0059] In this embodiment, the spherical distance refers to the shortest distance between two points on the spherical surface, and the calculation steps include:

[0060] Determine the geographic coordinates of two points: obtain the longitude and latitude coordinates of the two points on the earth's surface;

[0061] Convert longitude and latitude to radians: Most spherical distance calculations require longitude and latitude in radians, which is done by multiplying the angle by π / 180.

[0062] Apply the Haversine formula to calculate the great circle distance:

[0063]

[0064]

[0065] d=R·c

[0066] Among them, Δlat and Δlon represent the latitude difference and longitude difference between a certain track point of the vehicle trajectory and a corresponding track point of the mobile phone GPS, respectively. 1 and lat 2 are the latitudes of the vehicle trajectory point and the corresponding associated mobile phone trajectory point, respectively. R is the radius of the earth, and 6371 kilometers is usually used as the metric unit;

[0067] The calculated value d is the spherical distance between the two points.

[0068] In order to further implement the above technical solution, the specific content of threshold filtering is as follows:

[0069] Based on the daily vehicle code fitting calculation results, the scores of the first N accompanying mobile phone numbers of each vehicle are converted into scores based on z-score, and all mobile phone numbers with z-score scores lower than 1 are filtered out, where N is a preset value, usually 10;

[0070] Based on the z-score, the score is normalized and first-order difference is performed. The point with the largest difference value is taken as the cliff point, and the cliff point is the threshold. All results greater than the threshold are filtered out, and the remaining results are taken as the vehicle code fitting results after threshold filtering.

[0071] In this embodiment, the z-score calculation formula is:

[0072]

[0073] Among them, x is the original score, μ is the mean, and σ is the standard deviation. The significance of z-score is to standardize the original score, that is, to convert it into a relative position relative to the mean, so that it can be compared between different data distributions. By setting the threshold, a cliff can be formed between the scores, so as to further identify the accompanying numbers with scores significantly higher than the mean. By setting 1 as the cliff value, the accompanying numbers with z-score greater than 1 after the score conversion are found as cliff candidate numbers.

[0074] On this basis, the scores are normalized, then sorted in descending order, and the first-order difference calculation is performed on the top N scores;

[0075] Data normalization is to scale the data to a fixed range, usually between [0,1]. This normalization is usually called Min-Max normalization. Its advantage is that the original distribution shape of the data is maintained, and by scaling the data to the same range, the difference in data changes can be compared.

[0076] For a sequence {x i |i=1,…,n}, its first-order difference is defined as the difference between adjacent data points, which can be expressed as:

[0077] Δx i =x i+1 -x iThe first-order difference of a set of descending data can be used to extract cliff data. The principle is to use the difference value to identify the most significant change point in the data. According to the formula, the following conclusions can be drawn: the first-order difference is to calculate the difference between adjacent data points; the larger the difference value, the more significant the change between adjacent data points. Therefore, the maximum difference value usually represents a significant cliff in the data. Therefore, after the scores are normalized, they are sorted in descending order and the first-order difference is performed. The position with the largest difference value is the cliff position;

[0078] Finally, the result of taking out the cliff of z-score and the result of taking out the cliff after the first-order difference after normalization are taken into account, and the vehicle code fitting calculation result after threshold filtering is obtained.

[0079] In order to further implement the above technical solution, the specific content of sample filtering based on the first and last end trajectories is as follows:

[0080] Based on the vehicle code fitting calculation obtained by threshold filtering, all the trajectories with vehicle codes matching the timestamps of the preset proportion values ​​at the beginning and end in the marking results are marked, and the beginning and end trajectories are marked as 0 and 1 respectively, with the beginning being 0 and the end being 1;

[0081] The preset ratio value ranges from 5% to 15%. In this embodiment, the preset ratio value is set to 10%.

[0082] The vehicle and number are associated based on their 7-digit geo-hash timestamps, and all numbers that only match 0 or only match 1 are filtered out. Mobile phone numbers that match both 0 and 1 are retained as the high-level signal code set accompanying the vehicle code.

[0083] In order to further implement the above technical solution, the specific content of step S4 is: aggregate the high-confidence companion result candidate set of each vehicle above the threshold number of days on a monthly basis, and use the mobile phone number in the candidate set as the final vehicle code fitting mobile phone number.

[0084] In this embodiment, the threshold number of days is set to 3 days, that is, in the daily high confidence results, the numbers that appear for 3 days or more on a monthly basis are used as the mobile phone numbers for the final vehicle code fitting.

[0085] A vehicle code fitting system for highway scenarios, such as Figure 2 ,Based on a vehicle code fitting method for highway scenarios, it includes: a trajectory table creation module, a vehicle code fitting calculation module, a noise filtering module and an aggregation result output module;

[0086] The spatiotemporal trajectory table creation module is used to obtain highway vehicle trajectory data and mobile phone positioning data, extract key fields, and create vehicle spatiotemporal trajectory tables and mobile phone number spatiotemporal trajectory tables respectively;

[0087] The vehicle code fitting calculation module is used to perform vehicle code fitting calculation according to the vehicle time-space trajectory table and the mobile phone number time-space trajectory table to obtain the daily vehicle code fitting results;

[0088] The noise filtering module is used to reduce the noise of the daily vehicle code fitting results by using two types of noise filtering methods: threshold filtering and sample filtering based on the first and last trajectories to obtain a high-confidence accompanying result candidate set;

[0089] The result output module is used to aggregate the accompanying results of the final vehicle code fitting by month based on the high-confidence accompanying result candidate set.

[0090] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a vehicle code fitting method for a highway scene.

[0091] In this embodiment, the storage medium can be an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the above; more specific examples of storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, and any suitable combination of the above.

[0092] A processing terminal includes a memory and a processor. The memory stores a computer program that can be run on the processor. When the processor executes the computer program, a vehicle code fitting method for a highway scene is implemented.

[0093] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0094] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A vehicle code fitting method for highway scenes, characterized in that: The following steps are involved: S1. Obtain highway vehicle trajectory data and mobile phone positioning data, extract key fields, and create a vehicle spatiotemporal trajectory table and a mobile phone number spatiotemporal trajectory table respectively; S2. Perform vehicle code fitting calculation based on the vehicle time-space trajectory table and the mobile phone number time-space trajectory table to obtain daily fitting results; S3. Based on the daily fitting results, two types of noise filtering methods, threshold filtering and sample filtering based on the head and tail trajectories, are used to reduce the noise of the fitting results to obtain a high-confidence adjoint result candidate set; S4. Based on the high-confidence adjoint result candidate set, the final vehicle code fitting result is obtained by monthly aggregation; The specific contents of vehicle code fitting calculation include: S21. Based on the Haversine formula, the distance between the spatiotemporal trajectory point and the associated number trajectory point of each vehicle under each 7-bit timestamp and 5-bit geo-hash value is calculated day by day; S22. Find the closest mobile phone number trajectory point and the closest distance to each time and space trajectory point of each vehicle day by day; S23. The nearest distance is divided into bins according to the vehicle, the accompanying mobile phone number, the vehicle's 5-digit geographic hash value, and the 7-digit timestamp, and different bins correspond to different scores; S24. Aggregate the accompanying scores of all time and space points of each vehicle on a daily basis, and retain the top N accompanying mobile phone numbers with the highest scores on a single day, which is the daily vehicle code fitting calculation result; The specific content of step S4 is: aggregating the vehicle code fitting results of each vehicle above the threshold number of days on a monthly basis, and taking the candidate number set as the mobile phone number for the final vehicle code fitting.

2. The vehicle code fitting method for highway scenes according to claim 1 is characterized in that: The key fields of highway vehicle trajectory data include the vehicle unique identifier ID, the longitude of the highway checkpoint, the latitude of the highway checkpoint, the 12-bit geographic hash value of the highway checkpoint, and the timestamp and date of passing the current checkpoint.

3. The vehicle code fitting method for highway scenes according to claim 1 is characterized in that: The key fields of mobile phone location data include the unique identifier UID of the person to whom it belongs, the longitude of the mobile phone number track point, the latitude of the mobile phone number track point, the 12-digit geographic hash value of the mobile phone number track point, and the timestamp and date of the mobile phone number track point.

4. The vehicle code fitting method for highway scenes according to claim 1 is characterized in that: The specific contents of threshold filtering are: Based on the vehicle code fitting calculation results, the scores of the first N accompanying mobile phone numbers of each vehicle are converted into scores based on z-score every day, and all mobile phone numbers with z-score scores lower than 1 are filtered out; Based on the z-score, the score is normalized and first-order difference is performed. The point with the largest difference value is taken as the cliff point. All results after the cliff point are filtered, and all results before the cliff are retained as the vehicle code fitting results after threshold filtering.

5. The vehicle code fitting method for highway scenes according to claim 4 is characterized in that: The specific contents of sample filtering based on the first and last trajectories are as follows: Based on the vehicle code fitting result samples obtained by threshold filtering, the timestamps of all trajectories of the vehicle code accompanying matching pairs are marked as 0 and 1 respectively according to the preset proportion value, with the head end being 0 and the tail end being 1; The vehicle and number are associated based on their 7-digit geo-hash timestamps, and all numbers that only match 0 or only match 1 are filtered out. Mobile phone numbers that match both 0 and 1 are retained as the high-signal code set for vehicle code fitting.

6. A vehicle code fitting system for highway scenes, characterized in that: A vehicle code fitting method for a highway scene based on any one of claims 1 to 5, comprising: a trajectory table creation module, a vehicle code companion calculation module, a noise filtering module and an aggregation result output module; The spatiotemporal trajectory table creation module is used to obtain highway vehicle trajectory data and mobile phone positioning data, extract key fields respectively, and create vehicle spatiotemporal trajectory tables and mobile phone number spatiotemporal trajectory tables; The vehicle code accompanying calculation module is used to perform vehicle code fitting calculation according to the vehicle time-space trajectory table and the mobile phone number time-space trajectory table to obtain the daily vehicle code fitting results; The noise filtering module is used to reduce the noise of the daily companion fitting results by using two noise filtering methods: cliff threshold extraction filtering and sample filtering based on the head and tail trajectories, to obtain a high-confidence companion result candidate set; The result output module is used to obtain the final vehicle code fitting results by monthly aggregation based on the high-confidence accompanying result candidate set.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the vehicle code fitting method for a highway scene as described in any one of claims 1 to 5 is implemented.

8. A processing terminal, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the computer program, a vehicle code fitting method for a highway scene as described in any one of claims 1 to 5 is implemented.

Images