Traffic intersection safety warning method, system and storage medium

Abstract

This application provides a traffic intersection safety warning method, system, and storage medium. The method obtains intersection traffic data within a preset time window W (the preset time window W can be a time period including the current time), wherein the intersection traffic data includes at least intersection attribute data, intersection traffic volume data, intersection environment data, and data on the proportion of high-risk vehicles at the intersection. The intersection traffic data is then input into an intersection safety warning model S to obtain the hazard coefficient of the intersection. In summary, by observing the hazard coefficients of various intersections in an area, high-risk intersections can be identified in real time so that safety measures can be deployed at those intersections, thereby improving the efficiency of deploying manpower and material resources.

Description

Technical Field

[0001] The present application relates to the field of traffic safety, and specifically to a traffic intersection safety warning method, system and storage medium. Background Art

[0002] Traffic intersections play a crucial role in real life. Due to the high volume of vehicles and complex traffic conditions, the potential risk of traffic accidents at intersections increases dramatically, posing a significant safety hazard to people and property. Currently, two approaches are commonly used to mitigate risks at traffic intersections: one is to deploy traffic control personnel at intersections with relatively high traffic volumes based on historical experience, or to dispatch control personnel for on-site control upon receiving distress calls from certain vehicles. Both of these control methods are highly blind and lagging, resulting in a waste of manpower and material resources and low efficiency. Summary of the Invention

[0003] In order to solve the above technical problems, the technical solution adopted by the present application is: a traffic intersection safety warning method, comprising the following steps: S100, obtaining N traffic intersections C1, C2, ..., C N The intersection traffic data D within the set time window W is (D1, D2, ..., D N ), where the i-th traffic intersection C i The intersection traffic data D within the set time window W i At least including: intersection attribute data, intersection traffic flow data, intersection environment data, intersection high-risk vehicle ratio data, 1≤i≤N; S200, based on the intersection traffic data D and the intersection safety warning model S, obtain the N traffic intersections C1, C2, ..., C N The intersection risk factor Q=(Q1,Q2,...,Q N ), Q i C is the i-th traffic intersection i The intersection risk coefficient; S300, based on the intersection risk coefficient Q, the N traffic intersections C1, C2, ..., C N Provide safety warnings at high-risk intersections.

[0004] A traffic intersection safety warning system includes a processor and a non-transitory computer-readable storage medium, wherein the storage medium is used to store at least one instruction or at least one program, and is characterized in that the processor loads and executes the at least one instruction or at least one program to implement the above-mentioned method.

[0005] A computer-readable storage medium stores a program or instruction, wherein the program or instruction enables a computer to execute the steps of the above method.

[0006] The present application has at least the following technical effects: by acquiring intersection traffic data within a preset time window W (the preset time window can be a time period including the current time), wherein the intersection traffic data at least includes intersection attribute data, intersection traffic volume data, intersection environment data, and data on the proportion of high-risk vehicles at the intersection, and then inputting the intersection traffic data into an intersection safety warning model S, the hazard coefficient of the intersection can be obtained. That is, by observing the hazard coefficients of various intersections in the area, high-risk intersections can be identified in real time so that safety measures can be deployed at the intersection, preventing traffic accidents and improving the efficiency of deploying manpower and material resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0008] Figure 1 A flowchart of a traffic intersection safety warning method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0009] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0010] The present application embodiment provides a traffic intersection safety warning method, such as Figure 1 As shown, the method includes the following steps:

[0011] S100, obtain N traffic intersections C1, C2, ..., C in the first preset area. N The intersection traffic data D within the set time window W is (D1, D2, ..., D N ), where the i-th traffic intersection C i The intersection traffic data D within the set time window W i At least include: intersection attribute data, intersection traffic flow data, intersection environment data, and intersection high-risk vehicle ratio data, 1≤i≤N.

[0012] In the present application, the first preset area can be an area obtained in any manner, such as an administrative area, a geo-fenced area, etc. As an illustrative example, the administrative area can be a provincial-level administrative area or a prefecture-level administrative area, such as Hangzhou City in Zhejiang Province. A geo-fenced area is, for example, an area obtained by dividing the area according to longitude and latitude. The present application can select the first preset area according to actual needs. For example, the first preset area can be a district in a city or a city.

[0013] The N traffic intersections C1, C2, ..., C N It can be all traffic intersections within the first preset area, or it can be some traffic intersections within the first preset area. By setting N traffic intersections, the efficiency of safety warnings for intersections can be further improved, while reducing the workload of related data collection and calculation.

[0014] In this application, to better provide early warnings for intersections within a region, the duration of the set time window W is set within a range of [0.5 hours, 2 hours], preferably 1 hour. By setting the set time window W, the deployment of safety measures can be adjusted at any time according to actual needs, providing greater flexibility. Furthermore, the set time window W can be a historical time period, allowing for review of the traffic safety conditions at each intersection during that historical time period. The set time window W can also be a time period that includes the current time, allowing for review of real-time traffic conditions at each intersection and improving the efficiency of deploying real-time safety measures at each intersection.

[0015] Furthermore, the intersection attribute data is used to describe the intersection C i The objectively existing and unchangeable characteristics, in this application, the intersection attribute data at least include the following: intersection shape, importance level of all roads constituting the intersection, and the maximum importance level difference between all roads constituting the intersection. And the intersection attribute data in this application is not limited to the given content, and can also include any other data used to describe the attributes of the intersection. Specifically, the intersection shape, such as a crossroads, a Y-shaped intersection, a T-shaped intersection, etc., and the importance level of the roads constituting the intersection can be obtained from the prior art. An illustrative example is the classification of the levels of various roads in the process of urban development, such as national highways, provincial highways, rural roads, etc., and the classification of the importance of various roads in the city, such as main roads, ordinary roads, etc. When there are only two roads constituting a traffic intersection, and the importance level of one road is 1 and the other is 3, the maximum importance level difference between all roads constituting the intersection is 2.

[0016] The intersection traffic flow data is used to describe the intersection C iRegarding the traffic flow situation within the set time window W, in this application, the intersection traffic flow data at least includes the following contents: average traffic flow, traffic flow variation between intersections, traffic flow deviation, and average vehicle speed. It is known to those skilled in the art that the intersection traffic flow data may also include other data used to characterize the relevant characteristics of the intersection traffic flow. This application does not limit the content of the intersection traffic flow data. Specifically, the average traffic flow is used to represent the traffic flow of the intersection C. i The average of the unit traffic volume in a historical time. In one embodiment, the historical time is the most recent 24 hours including the current time. In this case, the average traffic volume = the sum of all unit traffic volumes in the most recent 24 hours / 24 hours. In another embodiment, the historical time is a specified day, then the average traffic volume = 24 unit traffic volumes in the specified day / 24 hours. The traffic flow deviation is used to represent the difference between the unit traffic volume of the intersection within the set time window W and the average traffic volume. For example, when the unit traffic volume is 100 and the average traffic volume is 120, it can be seen that the traffic flow deviation is 100-120=-20, wherein the unit traffic volume = the number of vehicles passing through the intersection C within the set time window W. i The average speed is the number of vehicles passing through the intersection C within the set time window W. i The average speed of all vehicles passing through the intersection C within the set time window W. i The sum of the speeds of all vehicles passing through the intersection C within the set time window W i The speed of a vehicle can be the speed collected when the vehicle passes through the middle of the intersection, or it can be the average speed of the vehicle passing through the intersection, for example, based on the distance / time traveled by the vehicle at the intersection. The flow change between intersections is used to represent the difference in the current unit traffic flow of adjacent intersections, for example, with intersection C i There are two adjacent intersections, and their current unit traffic volume is 100 and 50 respectively. i The current unit traffic volume is 100, and the traffic volume change between intersections is 0, 50.

[0017] The intersection environment data is used to describe the intersection C i The intersection environment data includes at least the weather conditions and the number of historical accidents at the intersection. In this application, the weather conditions may be, for example, sunny, cloudy, or thunderstorm weather, and the number of historical accidents at the intersection is the total number of accidents at the intersection found through a trusted data source, which may be some official data network.

[0018] The proportion of high-risk vehicles at the intersection = the number of vehicles passing through the intersection within the set time window W C i The number of high-risk vehicles passing through the traffic intersection C within the set time window W i Specifically, in the present application, the high-risk vehicles can be customized. For example, according to relevant management and control needs, vehicles with a history of traffic accidents exceeding a threshold value can be identified as high-risk vehicles. It is also possible to ignore the relevant management and control needs and identify vehicles with a history of traffic accidents exceeding another threshold value as high-risk vehicles.

[0019] S200, based on the intersection traffic data D and the intersection safety warning model S, obtain the N traffic intersections C1, C2, ..., C N The intersection risk factor Q=(Q1,Q2,...,Q N ), Q i C is the i-th traffic intersection i Specifically, in the present application, the intersection safety warning model S is a neural network model, such as BP, GBDT, random forest, and in the present application, the intersection safety warning model S is preferably a logistic regression model.

[0020] S300, based on the intersection risk coefficient Q, the N traffic intersections C1, C2, ..., C N Safety warnings are issued for high-risk intersections. In this step, high-risk warnings can be issued for the top few intersections with the highest risk factors. The specific number of warnings can be adaptively adjusted based on actual needs. Safety warnings can be issued using any of the existing methods, such as highlighting, jumping, or flashing displays.

[0021] As can be seen from the above, this application obtains intersection traffic data within a preset time window W (the preset time window can be a time period including the current time), where the intersection traffic data at least includes intersection attribute data, intersection traffic volume data, intersection environment data, and data on the proportion of high-risk vehicles at the intersection. The intersection traffic data is then input into the intersection safety warning model S to obtain the hazard coefficient of the intersection. In other words, by observing the hazard coefficients of various intersections in the area, high-risk intersections can be identified in real time so that safety measures can be deployed at those intersections, preventing traffic accidents and improving the efficiency of deploying manpower and material resources.

[0022] Furthermore, in this application, the acquisition of the intersection safety warning model S includes the following steps:

[0023] S201, obtain M traffic intersections R1, R2, ..., R in the second preset area. M In the preset time period P, G time units PU1, PU2, ..., PUG Traffic data of intersection within I=(I 11 ,I 12 ,...,I 1G ,I 21 ,I 22 ,...,I 2G ,...,I M1 ,I M2 ,...,I MG ) and intersection risk category data A=(A 11 ,A 12 ,...,A 1G ,A 21 ,A 22 ,...,A 2G ,...,A M1 ,A M2 ,...,A MG ), where the jth traffic intersection R j In the kth time unit PU k Traffic data of intersections within I jk At least include: intersection attribute data, intersection traffic flow data, intersection environment data, intersection high-risk vehicle ratio data, the jth traffic intersection R j In the kth time unit PU k The intersection risk category data is A jk , A jk ∈{0,1}, when the jth traffic intersection R j In the kth time unit PU k When the intersection is non-dangerous, A jk =0, otherwise A jk =1, 1≤j≤M, 1≤k≤G.

[0024] In this step, the second preset area can also be an area obtained in any way, such as an administrative area, a geographical fence area, etc., and the first preset area is included in the second preset area, that is, the first preset area can be the same as the second preset area, or the first preset area is part of the second preset area.

[0025] In one embodiment of the present application, the M traffic intersections R1, R2, ..., R M is part of the traffic intersections within the second preset area. At this time, the M traffic intersections R1, R2, ..., R M Contains the N access ports C1, C2, ..., C N In order to improve the robustness of the intersection safety warning model S, in a preferred embodiment of the present application, the M traffic intersections R1, R2, ..., R Mare all the traffic intersections within the second preset area. And from the above two embodiments, it can be seen that N≤M.

[0026] In order to obtain sufficient and comprehensive intersection traffic data, the preset time period P is in the range of [1 month, 3 months], preferably 1 month. k The duration range is [0.5 hours, 1 hour]. In one embodiment of the present application, the time unit PU k The duration of the preset time window W = the duration of the preset time window W = 1 hour. G = the duration of the preset time period P / time unit PU k For example, if the preset time period P is 1 month (the default is 30 days), and the time unit PU k For 1 hour, G = (24 / 1) * 30 = 720. It can be seen that at this time, all time units are 720 in total, and the time units do not intersect with each other, and the union of all time units is the preset time period P. In another embodiment of the present application, G < the duration of the preset time period P / time unit PU k duration.

[0027] S202, based on the intersection traffic data I=(I 11 ,I 12 ,...,I 1G ,I 21 ,I 22 ,...,I 2G ,...,I M1 ,I M2 ,...,I MG ) and intersection risk category data A=(A 11 ,A 12 ,...,A 1G ,A 21 ,A 22 ,...,A 2G ,...,A M1 ,A M2 ,...,A MG ) Train the logistic regression model to obtain the intersection safety warning model S.

[0028] Those skilled in the art will appreciate that, in order to effectively train the logistic regression model using the collected data, in this application, the collected intersection traffic data I is preprocessed so that the processed data can be directly input into the logistic regression model for training. Specifically, the preprocessing includes common feature digitization encoding and normalization processing. The feature digitization encoding method can adopt any method in the prior art. For example, for the intersection shapes in the acquired intersection attribute data, when the intersection shapes only include three types: crossroads, Y-shaped intersections, and T-shaped intersections, then the crossroads, Y-shaped intersections, and T-shaped intersections are digitized as 1, 2, and 3, respectively. Those skilled in the art will appreciate that any existing digital encoding method can be used to digitize different parts of the intersection traffic data. Normalization is a routine operation in the art and will not be elaborated here. In order to enable the use of the intersection safety warning model S, the intersection traffic data D adopts the same digitization encoding rules and normalization method as the intersection traffic data I.

[0029] From the above content, it can be seen that by collecting intersection traffic data of each traffic intersection in different time units within a preset historical time period in the second preset area including the first preset area, more comprehensive training data is obtained, which can fully improve the applicability and robustness of the intersection safety warning model S.

[0030] An embodiment of the present application further discloses a method for obtaining the traffic intersection C passing through the set time window W. i A method for determining the number of high-risk vehicles, the method comprising the following steps:

[0031] S101, obtaining the traffic intersection C within the set time window W i The vehicle identification data of all vehicles H=(H1,H2,...,H E ), where the vehicle identification data H of the e-th category vehicle e =(H e1 ,H e2 ,...,H ef(e) ), H eg is the vehicle identification of the gth vehicle in the eth category of vehicles, f(e) is the total number of vehicles in the eth category of vehicles, 1≤e≤E, 1≤g≤f(e), and E is the total number of types of all vehicles.

[0032] In this application, the vehicle identification of all vehicles can be obtained through the data collection device at the intersection. The vehicle identification is used to uniquely identify the vehicle, and the vehicle identification can be a license plate in the prior art. The data collection device is, for example, a road intersection photography device. The types of all vehicles can be divided into electric vehicles, private cars, motorcycles, buses, heavy vehicles, etc., so the value of E is related to the vehicle type classification. When it is divided only into electric vehicles, motorcycles and four-wheeled vehicles, E=3. When it is divided into electric vehicles, private cars, motorcycles, buses, and heavy vehicles, E=5.

[0033] S102, based on the vehicle identification data H, obtain the traffic driving data HD of all vehicles in a preset time period U = (HD1, HD2, ..., HD E ), where the traffic data HD of the e-th category vehicle e =(HD e1 ,HD e2 ,...,HD ef(e) ), the traffic driving data HD of the g-th vehicle in the e-th category eg At least including: list data of APPs installed on the vehicle driver’s mobile communication device, driver attribute data, driving habit data, and number of driving accidents.

[0034] In one embodiment of the present application, a list of apps installed on a vehicle driver's mobile communication device is stored in a first data source, which is a server. The server can exist in any manner known in the art, such as local storage, a cloud disk, an electronic device, a processor, etc. In this application, the form of the first data source is not limited. In another embodiment of the present application, the list of apps installed on the mobile communication device can be obtained by communicating with the vehicle driver's mobile communication device.

[0035] The driver attribute data and / or number of driving accidents in this application are derived from a second trusted data source, such as an official website, which stores at least all available vehicle identification information, driver-related information, vehicle driving data, number of vehicle accidents, and the time of vehicle accident occurrence. In this application, the driver attribute data includes at least: gender, age, education level, and occupation. The driver attribute data in this application may also include other data, such as company affiliation, etc. This application does not impose specific restrictions on the content of the driver attribute data.

[0036] Furthermore, the data in the second trusted data source can be accurately merged with the data in the first data source through a primary key value. In one embodiment, the data fusion between the first data source and the second trusted data source is performed through registered or recorded driver identification information, wherein the driver identification information can be, for example, an ID number or a phone number.

[0037] The driving habit data of the present application is used to describe the driver's driving habits, including at least: whether the driver drives for a long time, the proportion of nighttime driving, daytime speed, nighttime speed, and speed range. In this step, the collected data stored in the second trusted data source can be used to determine the driver's driving habits. For example, the driver's driving habit data can be inferred based on all the data collected from the driver during the preset time period U.

[0038] Specifically, the duration of the preset time period U ranges from [2 years to 5 years], preferably 3 years. By setting the duration of the preset time period U, as much and comprehensive data about the vehicle driver as possible can be obtained from the second trusted data source. Furthermore, the number of driving accidents is the number of accidents recorded by the vehicle driver during the preset time period U. This enhances the validity of the obtained data, preventing the applicability of subsequent high-risk vehicle identification models from being compromised due to outdated data.

[0039] In the present application, in order to enable the acquired data to directly train the model, the acquired feature data needs to be preprocessed, wherein the preprocessing includes digital coding and / or normalization, etc. Specifically, the digital coding can adopt any coding method in the prior art, such as the text-cnn method, the one-hot method, etc. For example, when the one-hot method is used for digital coding, for the "APP list data installed on the vehicle driver's mobile communication device", the first step is to obtain the total number of APPs that are safe on the mobile communication device of all vehicle drivers, and then use a sparse vector to represent each installed APP. For example, when the total number of APPs that are safe on the mobile communication device of all vehicle drivers is only 3, the first APP is coded as 100, the second APP is coded as 010, and the third APP is coded as 001. In the present application, other feature data in the HD can also be encoded using the one-hot method. However, those skilled in the art will understand that the digital coding method of the feature data can also adopt other methods.

[0040] S103, based on the traffic driving data HD and the high-risk vehicle judgment model SM=(SM1, SM2, ..., SM E ) Get the traffic intersection C within the set time window W iThe number of high-risk vehicles. Among them, based on traffic driving data HD e and the e-th category high-risk vehicle judgment model SM e To obtain the number of vehicles in category e that pass through the traffic intersection C within the set time window W i In this step, the traffic data of different types of vehicles are combined with the high-risk vehicle judgment model corresponding to their types, so as to obtain the number of high-risk vehicles in different vehicle types, and then obtain the number of vehicles passing through the traffic intersection C within the set time window W. i The number of high-risk vehicles of different types, where the time window W is set to pass through the traffic intersection C i The number of high-risk vehicles is the sum of the number of medium- and high-risk vehicles of different types.

[0041] From the above content, it can be seen that in order to identify the vehicles passing through the intersection C within the set time window W, i The number of high-risk vehicles passing through intersection C within the set time window W is collected in this application. i The vehicle identification data of all vehicles is combined with data from multiple data sources to obtain the vehicle driving data of all vehicles. The vehicle driving data at least includes the driver's Internet characteristics (installed APPs), driver attribute data, driver driving habit data, and the number of driving accidents; on this basis, the vehicle driving data of all vehicles are combined with the corresponding high-risk vehicle judgment model according to different types, and the number of high-risk vehicles in different types of vehicles is obtained. This application can further improve the accuracy of high-risk vehicle judgment by distinguishing different types of vehicles and setting high-risk judgment models for different types of vehicles.

[0042] In one embodiment of the present application, the e-th high-risk vehicle judgment model SM e The acquisition includes the following steps:

[0043] S1031, obtain M traffic intersections R1, R2, ..., R in the second preset area. M The historical traffic driving data DD of the e-th type of vehicles passing through the preset historical time period T e =(DD e1 ,DD e2 ,...,DD eh(e) ) and driving accident data B e =(B e1 ,B e2 ,...,B eh(e) ), where the historical traffic data DD of the fth vehicle in the eth category ef At least including: list of APPs installed on the vehicle driver's mobile communication device, driver attribute data, and driving habit data; B efis the number of driving accidents of the f-th vehicle in the e-th category, 1≤f≤h(e), h(e) is the number of M traffic intersections R1, R2, ..., R in the second preset area M The number of vehicles of category e that passed through the preset historical time period T.

[0044] In this application, as mentioned above, the historical traffic data DD can be obtained through the second trusted data source. e =(DD e1 ,DD e2 ,...,DD eh(e) ) and driving accident data B e =(B e1 ,B e2 ,...,B eh(e) In order to obtain more accurate and comprehensive data on vehicle drivers, the duration of T is in the range of [2 years, 5 years], preferably 3 years. The driving accident data is the total number of driving accidents within the preset historical time period T recorded by the second trusted data source.

[0045] In order to ensure data consistency, the duration of the preset historical time period T is the same as the duration of the preset time period U.

[0046] S1032, based on the historical traffic driving data DD e =(DD e1 ,DD e2 ,...,DD eh(e) ) and the driving accident data B e =(B e1 ,B e2 ,...,B eh(e) ) Obtain historical traffic data SD of high-risk vehicles e =(SD e1 ,SD e2 ,...,SD ep(e) ) and high-risk vehicle driving accident data SB e =(SB e1 ,SB e2 ,...,SB ep(e) ), where SD e For DD e The data corresponding to the number of driving accidents greater than the first risk driving threshold, SB em ∈B e And corresponds to SD em , 1≤m≤p(e), p(e) is SB eIn the present application, the value range of the first risk driving threshold is [3 times, 7 times], preferably 4 times. By adjusting the value of the first risk driving threshold, the deployment of safety measures at the intersection can be adjusted at any time according to the actual situation of the intersection.

[0047] S1033, based on the historical traffic data SD of the high-risk vehicle e =(SD e1 ,SD e2 ,...,SD ep(e) ) Perform cluster analysis to obtain V different grouping data SD1 e =(SD1 e1 ,SD1 e2 ,...,SD1 ex(e,1) ), SD2 e =(SD2 e1 ,SD2 e2 ,...,SD2 ex(e,2) ), ..., SDV e =(SDV e1 ,SDV e2 ,...,SDV ex(e,V) ), where the vth packet data SDv e The yth data SDv ey ∈SD e , 1≤y≤x(e,v), x(e,1)+x(e,2)+...+x(e,V)=p(e), 1≤v≤V. In this step, high-risk vehicles with a number of driving accidents greater than the first risk driving threshold are clustered and grouped, further obtaining a more refined classification of high-risk vehicles, facilitating more refined management and control of high-risk vehicles to meet changing management and control needs.

[0048] S1034, respectively based on SD1 e , SD2 e ,...,SDV e , train V logistic regression models to obtain the high-risk vehicle judgment model SM e Specifically, in this step, for the v-th logistic regression model, the v-th grouped data SDv e As positive sample data, and from the historical traffic driving data DD e Obtain historical traffic data SD of vehicles that are not high-risk e The x(e,v) data are used as negative sample data to train the v-th logistic regression model to obtain the high-risk vehicle judgment model SM e The vth high-risk vehicle judgment sub-model in the above equations can be deduced by analogy to obtain the high-risk vehicle judgment model SM. eSpecifically, the value of the grouping number V is 3 or 4, preferably 3.

[0049] From the above content, it can be seen that this application can carry out more refined management and control of high-risk vehicles by further clustering analysis of high-risk vehicles, making it easier to make targeted distinctions among high-risk vehicles according to management and control objectives, and then adopt different methods for education, management, etc.

[0050] Furthermore, in this application, based on the traffic driving data HD and the high-risk vehicle judgment model SM=(SM1, SM2, ..., SM E ) to obtain the traffic intersection C within the set time window W i The number of high-risk vehicles includes:

[0051] S31, based on the traffic driving data HD, obtain the first traffic driving sub-data HDU=(HDU1, HDU2, ..., HDU E ) and traffic driving second sub-data HDD=(HDD1, HDD2, ..., HDD E ), wherein the second sub-data HDD of the traffic driving of the g-th vehicle in the e-th category vehicle eg HD eg Number of driving accidents in HDD eg Corresponding HDU eg HD eg In other words, by putting the HDD eg Splicing in HDU eg HD available later eg .

[0052] S32, HDU1, HDU2, ..., HDU E Input high-risk vehicle judgment models SM1, SM2, ..., SM E To obtain the high-risk judgment coefficient QD of the vehicle, we need to calculate (QD1, QD2, ..., QD E ), where the high-risk judgment coefficient QD of the e-type vehicle is e =(QD e1 ,QD e2 ,...,QD ef(e) ), QD eg is the high-risk judgment coefficient of the g-th vehicle in the e-th category of vehicles. In one embodiment of the present application, the high-risk judgment coefficient QD of the g-th vehicle in the e-th category of vehicles is eg =∑ V v＝1 QD eg (v), where QD eg (v) To HDeg Input to high-risk vehicle judgment model SM e In the preferred embodiment of the present application, the high-risk judgment coefficient QD of the g-th vehicle in the e-th category is obtained by the v-th high-risk vehicle judgment sub-model. eg =max(QD eg (1),QD eg (2),...,QD eg (V)).

[0053] S33, based on the high-risk judgment coefficient QD of the vehicle = (QD1, QD2, ..., QD E ) and traffic driving second sub-data HDD=(HDD1, HDD2, ..., HDD E ) Obtaining the vehicle's location information in a preset vehicle risk table, wherein the preset vehicle risk table stores the vehicle identifiers corresponding to the vehicles in a preset sorting order. The preset sorting order is as follows: first, sorting the vehicles from largest to smallest by their second traffic travel sub-data; and when the second traffic travel sub-data of any two vehicles are identical, sorting the vehicles from largest to smallest by their high-risk determination coefficients. In the present application, the preset vehicle risk table is initially set to empty and is continuously updated as vehicle data is collected. As will be appreciated by those skilled in the art, the location information of any vehicle in the risk table will also change over time.

[0054] S34, based on the location information and risk control threshold of the vehicle in the preset vehicle risk table, obtain the number of vehicles passing through the traffic intersection C within the set time window W. i Specifically, in this step, when a vehicle's position in the preset vehicle risk table meets the risk control threshold, the vehicle is marked as a high-risk vehicle. The risk control threshold is set as a ratio threshold or a quantity threshold, which can be adjusted according to actual needs.

[0055] A traffic intersection safety warning system includes a processor and a non-transitory computer-readable storage medium, wherein the storage medium is used to store at least one instruction or at least one program. The processor loads and executes the at least one instruction or at least one program to implement the method provided in the above embodiment.

[0056] A computer-readable storage medium stores a program or instruction, wherein the program or instruction enables a computer to execute the steps of the method provided in the above embodiment.

[0057] An embodiment of the present application also provides a non-transitory computer-readable storage medium, which can be set in an electronic device to store at least one instruction or at least one program related to implementing a method in a method embodiment. The at least one instruction or the at least one program is loaded and executed by the processor to implement the method provided in the above embodiment.

[0058] An embodiment of the present application also provides an electronic device, including a processor and the aforementioned non-transitory computer-readable storage medium.

[0059] An embodiment of the present application further provides a computer program product comprising program code. When the program product is run on an electronic device, the program code is used to enable the electronic device to execute the steps of the method according to various exemplary embodiments of the present application described above in this specification.

[0060] Although some specific embodiments of the present application have been described in detail by way of example, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present application. It should also be understood by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the present application. The scope of the present application is defined by the appended claims.

Claims

1. A traffic intersection safety warning method, characterized in that: The following steps are involved: S100, obtain N traffic intersections C1, C2, ..., C in the first preset area. N The intersection traffic data D=(D1, D2, ..., D N ), Among them, the i-th traffic intersection C i The intersection traffic data D within the set time window W i At least including: intersection attribute data, intersection traffic flow data, intersection environment data, intersection high-risk vehicle ratio data, 1≤i≤N; the set time window W passes through the traffic intersection C i The number of high-risk vehicles is obtained by: S101, obtaining the traffic intersection C within the set time window W i The vehicle identification data of all vehicles H=(H1,H2,...,H E ), where the vehicle identification data H of the e-th category vehicle e =(H e1 ,H e2 ,...,H ef(e) ), H eg is the vehicle identifier of the g-th vehicle in the e-th category, f(e) is the total number of vehicles in the e-th category, 1≤e≤E, 1≤g≤f(e), and E is the total number of types of all vehicles; S102, based on the vehicle identification data H, obtain the traffic driving data HD of all vehicles in a preset time period U = (HD1, HD2, ..., HD E ), where the traffic data HD of the e-th category vehicle e =(HD e1 ,HD e2 ,...,HD ef(e) ), the traffic driving data HD of the g-th vehicle in the e-th category eg At least including: list of apps installed on the driver's mobile communication device, driver attribute data, driving habits data, and number of driving accidents; S103, based on the traffic data HD and the high-risk vehicle judgment model SM=(SM1, SM2, ..., SM E ) Get the traffic intersection C within the set time window W i The number of high-risk vehicles, including HD e and the e-th high-risk vehicle judgment model SM e To obtain the number of vehicles in category e that pass through the traffic intersection C within the set time window W i The number of high-risk vehicles in category E; Wherein, based on the traffic driving data HD and the high-risk vehicle judgment model SM=(SM1, SM2, ..., SM E ) to obtain the traffic intersection C within the set time window W i The number of high-risk vehicles includes: S31, based on the traffic driving data HD, obtain the first traffic driving sub-data HDU=(HDU1, HDU2, ..., HDU E ) and the second traffic driving sub-data HDD=(HDD1, HDD2, ..., HDD E ), wherein the second sub-data HDD of the traffic driving of the g-th vehicle in the e-th category vehicle eg HD eg Number of driving accidents in HDD eg Corresponding HDU eg HD eg Other data besides the number of driving accidents; S32, HDU1, HDU2, ..., HDU E Input high-risk vehicle judgment models SM1, SM2, ..., SM E To obtain the high-risk judgment coefficient QD of the vehicle = (QD1, QD2, ..., QD E ), where the high-risk judgment coefficient QD of the e-type vehicle is e =(QD e1 ,QD e2 ,...,QD ef(e) ), QD eg is the high-risk judgment coefficient of the g-th vehicle in the e-th category; S33, based on the vehicle's high-risk judgment coefficient QD=(QD1, QD2, ..., QD E ) and the second traffic driving sub-data HDD=(HDD1, HDD2, ..., HDD E ) obtaining position information of the vehicle in a preset vehicle risk table, wherein the preset vehicle risk table stores vehicle identifiers corresponding to the vehicles in a preset sorting order, wherein the preset sorting order is: first, sorting the vehicles from largest to smallest according to the second sub-data of their traffic movements, and when the second sub-data of the traffic movements of any two vehicles are the same, sorting the vehicles from largest to smallest according to their high-risk judgment coefficients; S34, based on the location information and risk control threshold of the vehicle in the preset vehicle risk table, obtain the number of vehicles passing through the traffic intersection C within the set time window W. i the number of high-risk vehicles; S200, based on the intersection traffic data D and the intersection safety warning model S, obtain the N traffic intersections C1, C2, ..., C N The intersection risk factor Q=(Q1,Q2,...,Q N ), Q i C is the i-th traffic intersection i The intersection hazard factor; S300, based on the intersection risk coefficient Q, the N traffic intersections C1, C2, ..., C N Provide safety warnings at high-risk intersections.

2. The safety warning method according to claim 1, characterized in that: The duration of the set time window W ranges from [0.5 hours to 2 hours].

3. The safety warning method according to claim 1 or 2, characterized in that: The intersection attribute data includes at least: the shape of the intersection, the importance level of all roads constituting the intersection, and the maximum importance level difference between all roads constituting the intersection; The intersection traffic flow data includes at least: average traffic flow, traffic flow variation between intersections, traffic flow deviation, and average vehicle speed; The intersection environment data includes at least: weather conditions, number of historical accidents at the intersection; The proportion of high-risk vehicles at the intersection = the number of vehicles passing through the intersection within the set time window W C i The number of high-risk vehicles passing through the traffic intersection C within the set time window W i The number of all vehicles.

4. The safety warning method according to claim 3, characterized in that: The acquisition of the intersection safety warning model S includes the following steps: S201, obtain M traffic intersections R1, R2, ..., R in the second preset area. M In the preset time period P, G time units PU1, PU2, ..., PU G Traffic data of intersections within I=(I 11 ,I 12 ,...,I 1G ,I 21 ,I 22 ,...,I 2G ,...,I M1 ,I M2 ,...,I MG ) and intersection risk category data A=(A 11 , A 12 ,...,A 1G ,A 21 ,A 22 ,...,A 2G ,...,A M1 ,A M2 ,...,A MG ), where the jth traffic intersection R j In the kth time unit PU k Traffic data of intersections within I jk At least include: intersection attribute data, intersection traffic flow data, intersection environment data, intersection high-risk vehicle ratio data, the jth traffic intersection R j In the kth time unit PU k The intersection risk category data is A jk , A jk ∈{0,1}, when the jth traffic intersection R j In the kth time unit PU k When the intersection is non-dangerous, A jk =0, otherwise A jk =1, 1≤j≤M, 1≤k≤G, traffic intersections R1, R2, ..., R M Including traffic intersections C1, C2, ..., C N , time unit PU k The duration of = the duration of the set time window W; S202 , training a logistic regression model based on the intersection traffic data I and the intersection risk category data A to obtain the intersection safety warning model S.

5. The safety warning method according to claim 4, characterized in that: The duration of the preset time period P ranges from [1 month, 3 months], and the time unit PU k The duration range is [0.5 hours, 1 hour].

6. The safety warning method according to claim 1, characterized in that: The e-th high-risk vehicle judgment model SM e The acquisition includes the following steps: S1031, obtain M traffic intersections R1, R2, ..., R in the second preset area. M The historical traffic driving data DD of the e-th type of vehicles passing through the preset historical time period T e =(DD e1 ,DD e2 ,...,DD eh(e) ) and driving accident data B e =(B e1 ,B e2 ,...,B eh(e) ), where the historical traffic data DD of the fth vehicle in the eth category ef At least including: list of APPs installed on the vehicle driver's mobile communication device, driver attribute data, and driving habit data; B ef is the number of driving accidents of the f-th vehicle in the e-th category within the preset historical time period T, 1≤f≤h(e), h(e) is the number of traffic intersections R1, R2, ..., R M The number of vehicles of category e that passed through a preset historical time period T, where the duration of the preset time period U = the duration of the preset historical time period T; S1032, based on the historical traffic driving data DD e =(DD e1 ,DD e2 ,...,DD eh(e) ) and the driving accident data B e =(B e1 ,B e2 ,...,B eh(e) ) Obtain historical traffic data SD of high-risk vehicles e =(SD e1 ,SD e2 ,...,SD ep(e) ) and high-risk vehicle driving accident data SB e =(SB e1 ,SB e2 ,...,SB ep(e) ), where SD e For DD e The data corresponding to the number of driving accidents greater than the first risk driving threshold, SB em ∈B e And corresponds to SD em , 1≤m≤p(e), p(e) is SB e The number of high-risk vehicles in the S1033, based on the historical traffic data SD of the high-risk vehicle e =(SD e1 ,SD e2 ,...,SD ep(e) ) Perform cluster analysis to obtain V different grouping data SD1 e =(SD1 e1 ,SD1 e2 ,...,SD1 ex(e,1) ), SD2 e =(SD2 e1 ,SD2 e2 ,...,SD2 ex(e,2) ), ..., SDV e =(SDV e1 ,SDV e2 ,...,SDV ex(e,V) ), where the vth packet data SDv e The yth data SDv ey ∈SD e , 1≤y≤x(e,v), x(e,1)+x(e,2)+...+x(e,V)=p(e), 1≤v≤V; S1034, respectively based on SD1 e , SD2 e ,...,SDV e , train V logistic regression models to obtain the high-risk vehicle judgment model SM e .

7. A traffic intersection safety warning system, characterized in that: The system includes a processor and a non-transitory computer-readable storage medium, which is used to store at least one instruction or at least one program, and is characterized in that the processor loads and executes the at least one instruction or at least one program to implement any one of claims 1-6.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program or instruction, and the program or instruction enables a computer to execute the steps of the method according to any one of claims 1 to 6.

Images