Road traffic remote monitoring and early warning system and method

Abstract

The invention discloses a road traffic remote monitoring and early warning system and method, and relates to the technical field of data analysis, and the method comprises the following steps: obtaining the information of a to-be-detected road section and the traffic accident data of the to-be-detected road section, carrying out the first combination of a second classification accident and a first accident probability, and obtaining a second classification accident; and performing second combination on the third accident classification and the comprehensive accident probability, establishing a second mapping relation among the first key feature, the second key feature and the comprehensive accident probability, obtaining the current comprehensive accident probability according to the first mapping relation, and performing early warning according to the comprehensive accident probability. According to the method, the traffic condition of the road section to be detected can be comprehensively known by integrating various types of data, potential risk factors of an accident can be more accurately identified through hierarchical classification and analysis, data support is provided for a traffic management department through analysis of a space-time distribution diagram, and the traffic safety is improved. Real-time monitoring and early warning can significantly improve the prevention ability of traffic accidents, reduce the accident rate, and improve the road safety.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a road traffic remote monitoring and early warning system and method. Background Art

[0002] In recent years, with the acceleration of urbanization and the frequent occurrence of traffic accidents, the research and application of traffic remote warning technology has gradually increased. With the continuous advancement of artificial intelligence and machine learning technology, the depth and breadth of data analysis will be improved, the intelligence level of warning will be enhanced, and big data analysis technology will be used to integrate data from different sources to identify traffic patterns and predict accident risks. Cloud computing provides powerful data processing capabilities and supports efficient data storage and analysis to reduce the accident rate, making traffic management more efficient and accurate.

[0003] At present, a road traffic accident early warning method is disclosed in a Chinese invention patent with publication number CN 106448149 B. The method combines high-risk drivers with high-accident sections, and adds the influence of environmental factors such as time and weather to achieve the purpose of judging high-accident drivers, judging the type of accidents that drivers have, and corresponding high-accident sections to specific driver types, and organically combining accident personnel, accident types, and accident sections. However, the relevant technology does not take into account the interference of sudden events, such as people breaking into and objects drifting in the way, on the probability of accident occurrence, which is easy to cause judgment errors, and does not establish a direct correspondence between the road's own attributes and human attributes and environmental attributes, resulting in excessive calculations, which is not conducive to the rapid responsiveness of the system and has certain limitations. Summary of the invention

[0004] The technical problem solved by the present invention is that the related technology does not take into account the interference of sudden events, such as people breaking in and objects drifting in the way, on the probability of accidents, which can easily lead to judgment errors. It does not establish a direct correspondence between the road's own attributes and the human attributes and environmental attributes, resulting in excessive calculations, which is not conducive to the rapid responsiveness of the system and has certain limitations.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: In the first aspect, a road traffic remote monitoring and early warning method comprises the following steps:

[0006] Step S100, obtaining information of a road section to be tested and traffic accident data of the road section to be tested, wherein the traffic accident data includes visible light intensity, ground friction, and accident-related monitoring video;

[0007] Step S200, setting visible light intensity and ground friction as first classification conditions, performing a first classification on traffic accident information according to the first classification conditions to obtain first classification accidents, extracting a first key feature of the accident-related monitoring video, analyzing interference causes according to the first key features, setting the interference causes as second classification conditions, performing a second classification on the first classification accidents according to the second classification conditions to obtain second classification accidents, counting the total number of traffic accidents on the road section to be tested, recorded as the first number, counting the number of second classification accidents, recorded as the second number, calculating the ratio of the first number to the second number, recorded as the first accident probability, and performing a first combination of the second classification accidents and the first accident probability;

[0008] Step S300, drawing a spatiotemporal distribution map of vehicle travel according to the accident-related monitoring video, extracting the second key feature of the spatiotemporal distribution map of vehicle travel, analyzing the cause of the accident according to the second key feature, setting the cause of the accident as a third classification condition, performing a third classification on the second classification accident to obtain a third classification accident, counting the number of third classification accidents, recording the third number, calculating the ratio of the third number to the first number, recording the second accident probability, calculating the comprehensive accident probability according to the second accident probability, and performing a second combination of the third accident classification and the comprehensive accident probability;

[0009] Step S400, establishing a second mapping relationship between the first key feature, the second key feature and the comprehensive accident probability, obtaining the current comprehensive accident probability according to the first mapping relationship, and issuing an early warning according to the comprehensive accident probability.

[0010] As a preferred solution of the road traffic remote monitoring and early warning method described in the present invention, step S100 includes the following sub-steps:

[0011] Step S101: number the road sections to be tested, where the numbers are represented by X i , where i is a natural number;

[0012] Step S102, obtaining information of the road section to be tested and traffic accident data of the road section to be tested, wherein the information of the road section to be tested includes traffic density, road inclination angle and road curvature, and the traffic accident data includes visible light intensity, ground friction and accident-related monitoring video, and the accident-related monitoring video is represented by the monitoring video corresponding to the first time period before the time point of the accident occurrence, and the first time period is obtained through big data analysis.

[0013] As a preferred solution of the road traffic remote monitoring and early warning method described in the present invention, step S200 includes the following sub-steps:

[0014] Step S201, setting visible light and ground friction as first classification conditions, setting a first value as a visible light threshold, and setting a second value as a ground friction threshold, performing a first classification on traffic accident information according to the first classification conditions to obtain first classification accidents, wherein the first classification accidents include first classification accidents in haze weather, first classification accidents in non-haze weather, first classification accidents in rainy and snowy weather, and first classification accidents in sunny weather;

[0015] Step S202, performing frame segmentation and windowing processing on the accident-related surveillance video to obtain a surveillance image, and extracting a first key feature of the surveillance image, wherein the first key feature is a graphic feature, and the first key feature includes a person, a cat, a dog, a plastic bag, a leaf, and air;

[0016] Step S203: analyzing the interference cause according to the first key feature, wherein the analysis logic of the interference cause includes:

[0017] Obtain the first key feature. When the first key feature is air, set the interference cause to no interference. When the first key feature is a person, a cat, or a dog, set the interference cause to animal interference. When the first key feature is a plastic bag or a leaf, set the interference cause to still object interference.

[0018] Step S204, setting the interference cause as a second classification condition, and performing a second classification on the first classification accident according to the second classification condition to obtain a second classification accident, wherein the second classification accident includes an animal interference second classification accident and a static interference second classification accident, and the second classification accident corresponds to the first classification accident;

[0019] Step S205, counting the total number of traffic accidents on the road section to be tested, recorded as the first number, counting the number of second-category accidents, recorded as the second number, calculating the ratio of the first number to the second number, recorded as the first accident probability, and performing a first combination of the second-category accidents and the first accident probability.

[0020] As a preferred solution of the road traffic remote monitoring and early warning method described in the present invention, the classification logic of the first classification accident includes:

[0021] Obtain visible light and ground friction, set a first value as a visible light threshold, and set a second value as a ground friction threshold, wherein the first value and the second value are obtained through big data analysis, and according to historical meteorological information, respectively calculate an average value of visible light corresponding to haze weather and an average value of ground friction corresponding to rainy and snowy weather, and set the average value of visible light corresponding to haze weather as the visible light threshold, and set the average value of ground friction corresponding to rainy and snowy weather as the ground friction threshold;

[0022] Compare the visible light with the first value, and when the visible light is less than or equal to the first value, set the first category accident as a first category accident in haze weather; when the visible light is greater than the first value, set the first category accident as a first category accident in non-haze weather;

[0023] The ground friction force is compared with the second value. When the ground friction force is greater than or equal to the second value, the first category accident is set as a first category accident in rainy and snowy weather. When the ground friction force is less than the second value, the first category accident is set as a first category accident in sunny weather.

[0024] Among them, the first category accidents related to visible light and the first category accidents related to ground friction can be superimposed.

[0025] As a preferred solution of the road traffic remote monitoring and early warning method described in the present invention, the extraction logic of the first key feature includes:

[0026] Call the graphic database, select any picture in the graphic database, record it as the standard picture, extract the graphic feature of the standard picture, obtain the monitoring image, extract the graphic feature of the monitoring image, calculate the similarity between the graphic feature of the monitoring image and the graphic feature of the standard picture by the cosine similarity formula, traverse the similarity, sort the similarity in descending order, select the standard picture corresponding to the similarity with the largest value, set it as the first key image, set the graphic feature corresponding to the first key image as the first key feature, set the similarity threshold, compare the similarity with the similarity threshold, and when all similarities are less than the similarity threshold, set the first key feature to air.

[0027] As a preferred solution of the road traffic remote monitoring and early warning method described in the present invention, step S300 includes the following sub-steps:

[0028] Step S301, obtain a monitoring image, obtain a time point corresponding to the monitoring image, record it as a key point, track the accident vehicle before the accident according to the accident vehicle license plate, match the position coordinates of the accident vehicle at the key point, the position coordinates are expressed as the coordinates of the center of gravity of the accident vehicle, set the position coordinates as the vehicle trajectory point, calculate the speed of the accident vehicle, and calculate the average value of the speed of the accident vehicle between the key points, set the average value of the speed of the accident vehicle between the key points as the average speed of the vehicle, set a horizontal road surface baseline, obtain the angle between the front of the accident vehicle and the horizontal road surface baseline, and calculate the average value of the angle between the key points, set the average value of the angle between the key points as the average steering angle of the vehicle, and draw a time-space distribution diagram of vehicle travel;

[0029] The vehicle travel time-space distribution diagram includes vehicle track points, vehicle track curves, vehicle track direction arrows, vehicle average speed and vehicle average steering angle;

[0030] Step S302, extracting a second key feature of the vehicle travel time-space distribution diagram, where the second key feature is a numerical feature, analyzing the cause of the accident according to the second key feature, setting the cause of the accident as a third classification condition, performing a third classification on the second classification accident, and obtaining a third classification accident;

[0031] The third category accidents correspond to the second category accidents, and the third category accidents include overspeed third category accidents, oversteering third category accidents, normal speed third category accidents and normal steering category accidents. The steering-related third category accidents and speed-related third category accidents can be superimposed.

[0032] Step S303, counting the number of accidents of the third category, recording it as the third number, calculating the ratio of the third number to the first number, recording it as the second accident probability, calculating the product of the second accident probability and the first accident probability, recording it as the comprehensive accident probability, and performing a second combination of the third accident category and the comprehensive accident probability;

[0033] The vehicle trajectory points are obtained according to the position coordinates and the point drawing method. The vehicle trajectory curve is obtained by inserting n supplementary points between adjacent trajectory points through the interpolation method and connecting them through a smooth curve. The adjacent trajectory points are represented as position coordinate points corresponding to the key points in the time sequence. The smooth curve is represented as a function corresponding to the vehicle trajectory curve having a first-order continuous derivative;

[0034] The extraction logic of the second key feature includes:

[0035] Obtaining an average vehicle speed and an average vehicle steering angle, setting a third value as a vehicle average speed threshold, and setting a fourth value as a vehicle average steering angle threshold, wherein the third value and the fourth value are obtained through big data analysis;

[0036] The average speed of the vehicle is compared with the third value, and when the average speed of the vehicle is greater than the third value, the second key feature is set to be too fast, otherwise, the second key feature is set to be normal;

[0037] The average steering angle of the vehicle is compared with the fourth value. When the average steering angle of the vehicle is greater than the fourth value, the second feature is set to excessive steering; otherwise, the second key feature is set to normal steering.

[0038] As a preferred solution of the road traffic remote monitoring and early warning method described in the present invention, step S400 includes the following sub-steps:

[0039] Step S401, establishing a second mapping relationship between the first key feature, the second key feature and the comprehensive accident probability;

[0040] Step S402, obtaining the current surveillance video, extracting the current second key feature and the current first key feature according to the current surveillance video, performing a first match between the current first key feature and the first key feature, and performing a second match between the current second key feature and the second key feature, wherein the first match of the current first key feature is the same as the extraction logic of the first key feature, and the second match of the current second key feature is the same as the extraction logic of the second key feature;

[0041] Step S403, obtaining a comprehensive accident probability corresponding to the result of the first match and the result of the second match, sending a first warning signal according to the comprehensive accident probability, and setting a fifth value and a sixth value as a probability threshold, wherein the fifth value and the sixth value are both distributed between 0 and 1, and the fifth value is less than the sixth value;

[0042] When the comprehensive accident probability is less than or equal to the fifth value, a warning signal of the first brightness and the first color is issued, and the corresponding comprehensive accident probability is displayed; when the comprehensive accident probability is greater than the fifth value and less than or equal to the sixth value, a warning signal of the second brightness and the second color is issued, and the corresponding comprehensive accident probability is displayed; when the comprehensive accident probability is greater than the sixth value, a warning signal of the third brightness and the third color is issued, and the corresponding comprehensive accident probability is displayed, wherein the corresponding comprehensive accident probability is displayed through the OLED display screen.

[0043] In a second aspect, a road traffic remote monitoring and early warning system includes a collection module, an analysis module and an early warning module;

[0044] The acquisition module obtains information of the road section to be tested and traffic accident data of the road section to be tested, wherein the traffic accident data includes visible light intensity, ground friction and accident-related monitoring video;

[0045] The analysis module sets visible light intensity and ground friction as the first classification condition, performs a first classification on the traffic accident information according to the first classification condition to obtain a first classification accident, extracts a first key feature of the accident-related monitoring video, analyzes the interference cause according to the first key feature, sets the interference cause as the second classification condition, performs a second classification on the first classification accident according to the second classification condition to obtain a second classification accident, counts the total number of traffic accidents on the road section to be tested, recorded as the first number, counts the number of second classification accidents, recorded as the second number, calculates the ratio of the first number to the second number, recorded as the first accident probability, and performs a first combination of the second classification accident and the first accident probability, draws a vehicle travel time-space distribution map according to the accident-related monitoring video, extracts the second key feature of the vehicle travel time-space distribution map, analyzes the cause of the accident according to the second key feature, sets the cause of the accident as the third classification condition, performs a third classification on the second classification accident, obtains a third classification accident, counts the number of third classification accidents, recorded as the third number, calculates the ratio of the third number to the first number, recorded as the second accident probability, calculates the comprehensive accident probability according to the second accident probability, and performs a second combination of the third accident classification and the comprehensive accident probability;

[0046] The early warning module establishes a second mapping relationship between the first key feature, the second key feature and the comprehensive accident probability, obtains the current comprehensive accident probability according to the first mapping relationship, and issues an early warning according to the comprehensive accident probability.

[0047] In a third aspect, the present invention provides an electronic device comprising a memory, a processor and storage in the memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps in any one of the methods described above are executed.

[0048] In a fourth aspect, the present invention provides a storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps in any one of the methods described above are executed.

[0049] The beneficial effects of the present invention are as follows: by integrating various types of data, it is possible to fully understand the traffic conditions of the road section to be tested and ensure the accuracy of subsequent analysis; through this hierarchical classification and analysis, the potential risk factors of accidents can be more accurately identified, providing a basis for subsequent accident warnings; through the analysis of the spatiotemporal distribution map, the high-incidence areas of accidents at specific times and locations can be identified, thereby providing data support for traffic management departments and improving traffic safety; real-time monitoring and warning can significantly improve the ability to prevent traffic accidents, reduce the accident rate, and improve road safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A basic flow chart of a road traffic remote monitoring and early warning method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0052] Example, see Figure 1 , as an embodiment of the present invention, provides a road traffic remote monitoring and early warning method, comprising the following steps:

[0053] Step S100, obtaining information of a road section to be tested and traffic accident data of the road section to be tested, wherein the traffic accident data includes visible light intensity, ground friction, and accident-related monitoring video;

[0054] Step S200, setting visible light intensity and ground friction as first classification conditions, performing a first classification on traffic accident information according to the first classification conditions to obtain first classification accidents, extracting a first key feature of the accident-related monitoring video, analyzing interference causes according to the first key features, setting the interference causes as second classification conditions, performing a second classification on the first classification accidents according to the second classification conditions to obtain second classification accidents, counting the total number of traffic accidents on the road section to be tested, recorded as the first number, counting the number of second classification accidents, recorded as the second number, calculating the ratio of the first number to the second number, recorded as the first accident probability, and performing a first combination of the second classification accidents and the first accident probability;

[0055] Step S300, drawing a spatiotemporal distribution map of vehicle travel according to the accident-related monitoring video, extracting the second key feature of the spatiotemporal distribution map of vehicle travel, analyzing the cause of the accident according to the second key feature, setting the cause of the accident as a third classification condition, performing a third classification on the second classification accident to obtain a third classification accident, counting the number of third classification accidents, recording the third number, calculating the ratio of the third number to the first number, recording the second accident probability, calculating the comprehensive accident probability according to the second accident probability, and performing a second combination of the third accident classification and the comprehensive accident probability;

[0056] Step S400, establishing a second mapping relationship between the first key feature, the second key feature and the comprehensive accident probability, obtaining the current comprehensive accident probability according to the first mapping relationship, and issuing an early warning according to the comprehensive accident probability.

[0057] By integrating various types of data, the present invention can fully understand the traffic conditions of the road section to be tested and ensure the accuracy of subsequent analysis. Through this hierarchical classification and analysis, the potential risk factors of accidents can be identified more accurately, providing a basis for subsequent accident warnings. Through the analysis of spatiotemporal distribution maps, high-incidence areas of accidents at specific times and locations can be identified, thereby providing data support for traffic management departments and improving traffic safety. Real-time monitoring and warning can significantly improve the ability to prevent traffic accidents, reduce the accident rate, and improve road safety.

[0058] The step S100 includes the following sub-steps:

[0059] Step S101, numbering the road sections to be tested, the numbering is represented by Xi, where i is a natural number;

[0060] Step S102, obtaining information of the road section to be tested and traffic accident data of the road section to be tested, wherein the information of the road section to be tested includes traffic density, road inclination angle and road curvature, and the traffic accident data includes visible light intensity, ground friction and accident-related monitoring video, and the accident-related monitoring video is represented by the monitoring video corresponding to the first time period before the time point of the accident occurrence, and the first time period is obtained through big data analysis.

[0061] In specific implementation, by assigning a unique number to each section to be tested, subsequent data management and analysis will be facilitated, ensuring that the information of each section can be accurately associated with the accident data, which will help improve the operability and efficiency of the system. By integrating various information, a comprehensive assessment of the section to be tested can be formed, ensuring that subsequent accident analysis and early warning measures are more scientific and effective. By using data such as visible light and ground friction, combined with accident monitoring videos, the environmental factors and direct causes of accidents can be better analyzed, thereby improving the effectiveness of accident prevention.

[0062] The step S200 includes the following sub-steps:

[0063] Step S201, setting visible light and ground friction as first classification conditions, setting a first value as a visible light threshold, and setting a second value as a ground friction threshold, performing a first classification on traffic accident information according to the first classification conditions to obtain first classification accidents, wherein the first classification accidents include first classification accidents in haze weather, first classification accidents in non-haze weather, first classification accidents in rainy and snowy weather, and first classification accidents in sunny weather;

[0064] Step S202, performing frame segmentation and windowing processing on the accident-related surveillance video to obtain a surveillance image, and extracting a first key feature of the surveillance image, wherein the first key feature is a graphic feature, and the first key feature includes a person, a cat, a dog, a plastic bag, a leaf, and air;

[0065] Step S203: analyzing the interference cause according to the first key feature, wherein the analysis logic of the interference cause includes:

[0066] Obtain the first key feature. When the first key feature is air, set the interference cause to no interference. When the first key feature is a person, a cat, or a dog, set the interference cause to animal interference. When the first key feature is a plastic bag or a leaf, set the interference cause to still object interference.

[0067] Step S204, setting the interference cause as a second classification condition, and performing a second classification on the first classification accident according to the second classification condition to obtain a second classification accident, wherein the second classification accident includes an animal interference second classification accident and a static interference second classification accident, and the second classification accident corresponds to the first classification accident;

[0068] Step S205, counting the total number of traffic accidents on the road section to be tested, recorded as the first number, counting the number of second-category accidents, recorded as the second number, calculating the ratio of the first number to the second number, recorded as the first accident probability, and performing a first combination of the second-category accidents and the first accident probability.

[0069] In specific implementation, by classifying accident information according to weather conditions and road conditions, the environmental factors that cause accidents can be better identified and analyzed, which is convenient for subsequent risk assessment and early warning measures. By processing and extracting features from surveillance videos, interference objects related to accidents can be more accurately identified, which helps to analyze the specific circumstances when the accident occurred. By classifying the causes of interference, the background of the accident can be more deeply understood, and external factors that may have led to the accident can be identified, providing guidance for accident prevention measures. Secondary classification can further refine the types of accidents and provide a clearer analysis perspective, so as to more effectively formulate targeted safety intervention measures. By counting the number of accidents and calculating the ratio, the probability of occurrence of different types of accidents can be quantified, providing a scientific basis for traffic safety management.

[0070] The classification logic of the first classification accident includes:

[0071] Obtain visible light and ground friction, set a first value as a visible light threshold, and set a second value as a ground friction threshold, wherein the first value and the second value are obtained through big data analysis, and according to historical meteorological information, respectively calculate an average value of visible light corresponding to haze weather and an average value of ground friction corresponding to rainy and snowy weather, and set the average value of visible light corresponding to haze weather as the visible light threshold, and set the average value of ground friction corresponding to rainy and snowy weather as the ground friction threshold;

[0072] Compare the visible light with the first value, and when the visible light is less than or equal to the first value, set the first category accident as a first category accident in haze weather; when the visible light is greater than the first value, set the first category accident as a first category accident in non-haze weather;

[0073] The ground friction force is compared with the second value. When the ground friction force is greater than or equal to the second value, the first category accident is set as a first category accident in rainy and snowy weather. When the ground friction force is less than the second value, the first category accident is set as a first category accident in sunny weather.

[0074] Among them, the first category accidents related to visible light and the first category accidents related to ground friction can be superimposed.

[0075] In specific implementation, historical data are used to set thresholds to improve the scientificity and accuracy of accident classification and ensure that the classification standards can effectively reflect the actual situation. Through big data analysis, key environmental factors under specific weather conditions can be identified, providing solid data support for accident analysis. By comparing the actual value of visible light with the set threshold, accidents caused by haze weather can be effectively identified, helping traffic management departments to strengthen traffic control and safety reminders under severe weather conditions. Superimposed classification can provide more comprehensive information for accident analysis. For example, in rainy and snowy weather with low visible light, a higher accident rate may occur.

[0076] The extraction logic of the first key feature includes:

[0077] Call the graphic database, select any picture in the graphic database, record it as the standard picture, extract the graphic feature of the standard picture, obtain the monitoring image, extract the graphic feature of the monitoring image, calculate the similarity between the graphic feature of the monitoring image and the graphic feature of the standard picture by the cosine similarity formula, traverse the similarity, sort the similarity in descending order, select the standard picture corresponding to the similarity with the largest value, set it as the first key image, set the graphic feature corresponding to the first key image as the first key feature, set the similarity threshold, compare the similarity with the similarity threshold, and when all similarities are less than the similarity threshold, set the first key feature to air.

[0078] In the specific implementation, a standard image is selected as the benchmark for feature extraction to ensure the accuracy of subsequent similarity calculations. By extracting the features of the surveillance image, necessary information is provided for subsequent similarity calculations to ensure that important elements in the image can be effectively identified. The cosine similarity quantifies the similarity between two feature vectors. The mechanism of setting the first key feature as air helps to simplify interference analysis in the absence of obvious interfering objects and ensure the validity of the results.

[0079] The step S300 includes the following sub-steps:

[0080] Step S301, obtain a monitoring image, obtain a time point corresponding to the monitoring image, record it as a key point, track the accident vehicle before the accident according to the accident vehicle license plate, match the position coordinates of the accident vehicle at the key point, the position coordinates are expressed as the coordinates of the center of gravity of the accident vehicle, set the position coordinates as the vehicle trajectory point, calculate the speed of the accident vehicle, and calculate the average value of the speed of the accident vehicle between the key points, set the average value of the speed of the accident vehicle between the key points as the average speed of the vehicle, set a horizontal road surface baseline, obtain the angle between the front of the accident vehicle and the horizontal road surface baseline, and calculate the average value of the angle between the key points, set the average value of the angle between the key points as the average steering angle of the vehicle, and draw a time-space distribution diagram of vehicle travel;

[0081] The vehicle travel time-space distribution diagram includes vehicle track points, vehicle track curves, vehicle track direction arrows, vehicle average speed and vehicle average steering angle;

[0082] Step S302, extracting a second key feature of the vehicle travel time-space distribution diagram, where the second key feature is a numerical feature, analyzing the cause of the accident according to the second key feature, setting the cause of the accident as a third classification condition, performing a third classification on the second classification accident, and obtaining a third classification accident;

[0083] The third category accidents correspond to the second category accidents, and the third category accidents include overspeed third category accidents, oversteering third category accidents, normal speed third category accidents and normal steering category accidents. The steering-related third category accidents and speed-related third category accidents can be superimposed.

[0084] Step S303, counting the number of accidents of the third category, recording it as the third number, calculating the ratio of the third number to the first number, recording it as the second accident probability, calculating the product of the second accident probability and the first accident probability, recording it as the comprehensive accident probability, and performing a second combination of the third accident category and the comprehensive accident probability;

[0085] The vehicle trajectory points are obtained according to the position coordinates and the point drawing method. The vehicle trajectory curve is obtained by inserting n supplementary points between adjacent trajectory points through the interpolation method and connecting them through a smooth curve. The adjacent trajectory points are represented as position coordinate points corresponding to the key points in the time sequence. The smooth curve is represented as a function corresponding to the vehicle trajectory curve having a first-order continuous derivative;

[0086] The extraction logic of the second key feature includes:

[0087] Obtaining an average vehicle speed and an average vehicle steering angle, setting a third value as a vehicle average speed threshold, and setting a fourth value as a vehicle average steering angle threshold, wherein the third value and the fourth value are obtained through big data analysis;

[0088] The average speed of the vehicle is compared with the third value, and when the average speed of the vehicle is greater than the third value, the second key feature is set to be too fast, otherwise, the second key feature is set to be normal;

[0089] The average steering angle of the vehicle is compared with the fourth value. When the average steering angle of the vehicle is greater than the fourth value, the second feature is set to excessive steering; otherwise, the second key feature is set to normal steering.

[0090] In specific implementation, accurate calculation of speed and steering angle can provide more accurate feature data for accident analysis, and the cause of the accident can be combined with speed and steering angle to achieve multi-level accident classification, which can more effectively identify the cause of the accident. Through the thresholds obtained through big data analysis, data-driven accident prevention and intervention measures can be implemented to improve the scientificity and effectiveness of traffic management. The drawing of the spatiotemporal distribution map of vehicle driving helps to intuitively understand the vehicle driving conditions and potential dangers, thereby providing visual support for accident analysis and prevention. Through the calculation of comprehensive accident probability, the accident risk can be evaluated, which helps to optimize traffic management strategies.

[0091] The step S400 includes the following sub-steps:

[0092] Step S401, establishing a second mapping relationship between the first key feature, the second key feature and the comprehensive accident probability;

[0093] Step S402, obtaining the current surveillance video, extracting the current second key feature and the current first key feature according to the current surveillance video, performing a first match between the current first key feature and the first key feature, and performing a second match between the current second key feature and the second key feature, wherein the first match of the current first key feature is the same as the extraction logic of the first key feature, and the second match of the current second key feature is the same as the extraction logic of the second key feature;

[0094] Step S403, obtaining a comprehensive accident probability corresponding to the result of the first match and the result of the second match, sending a first warning signal according to the comprehensive accident probability, and setting a fifth value and a sixth value as a probability threshold, wherein the fifth value and the sixth value are both distributed between 0 and 1, and the fifth value is less than the sixth value;

[0095] When the comprehensive accident probability is less than or equal to the fifth value, a warning signal of the first brightness and the first color is issued, and the corresponding comprehensive accident probability is displayed; when the comprehensive accident probability is greater than the fifth value and less than or equal to the sixth value, a warning signal of the second brightness and the second color is issued, and the corresponding comprehensive accident probability is displayed; when the comprehensive accident probability is greater than the sixth value, a warning signal of the third brightness and the third color is issued, and the corresponding comprehensive accident probability is displayed, wherein the corresponding comprehensive accident probability is displayed through the OLED display screen.

[0096] In specific implementation, the probability-based warning mechanism can provide timely feedback to help relevant personnel respond quickly. Warning signals of different brightness and colors provide intuitive identification of danger levels, which helps to improve the effectiveness of traffic safety management. The use of OLED display screens can ensure that information is clearly visible and enhance the user experience of the warning system.

[0097] By integrating various types of data, the present invention can fully understand the traffic conditions of the road section to be tested and ensure the accuracy of subsequent analysis. Through this hierarchical classification and analysis, the potential risk factors of accidents can be identified more accurately, providing a basis for subsequent accident warnings. Through the analysis of spatiotemporal distribution maps, high-incidence areas of accidents at specific times and locations can be identified, thereby providing data support for traffic management departments and improving traffic safety. Real-time monitoring and warning can significantly improve the ability to prevent traffic accidents, reduce the accident rate, and improve road safety.

[0098] It should be understood by those skilled in the art that the embodiments of the present invention can be provided as methods, systems or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program codes. Among them, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, disk or optical disk. These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0099] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A road traffic remote monitoring and early warning method, characterized in that: The following steps are involved: Step S100, obtaining information of a road section to be tested and traffic accident data of the road section to be tested, wherein the traffic accident data includes visible light intensity, ground friction, and accident-related monitoring video; Step S200, setting visible light intensity and ground friction as first classification conditions, performing a first classification on traffic accident information according to the first classification conditions to obtain first classification accidents, extracting a first key feature of the accident-related monitoring video, analyzing interference causes according to the first key features, setting the interference causes as second classification conditions, performing a second classification on the first classification accidents according to the second classification conditions to obtain second classification accidents, counting the total number of traffic accidents on the road section to be tested, recorded as the first number, counting the number of second classification accidents, recorded as the second number, calculating the ratio of the first number to the second number, recorded as the first accident probability, and performing a first combination of the second classification accidents and the first accident probability; Step S300, drawing a spatiotemporal distribution map of vehicle travel according to the accident-related monitoring video, extracting the second key feature of the spatiotemporal distribution map of vehicle travel, analyzing the cause of the accident according to the second key feature, setting the cause of the accident as a third classification condition, performing a third classification on the second classification accident to obtain a third classification accident, counting the number of third classification accidents, recording the third number, calculating the ratio of the third number to the first number, recording the second accident probability, calculating the comprehensive accident probability according to the second accident probability, and performing a second combination of the third accident classification and the comprehensive accident probability; Step S400, establishing a second mapping relationship between the first key feature, the second key feature and the comprehensive accident probability, obtaining the current comprehensive accident probability according to the first mapping relationship, and issuing an early warning according to the comprehensive accident probability.

2. A road traffic remote monitoring and early warning method as claimed in claim 1, characterized in that: The step S100 includes the following sub-steps: Step S101: number the road sections to be tested, where the numbers are represented by X i , where i is a natural number; Step S102, obtaining information of the road section to be tested and traffic accident data of the road section to be tested, wherein the information of the road section to be tested includes traffic density, road inclination angle and road curvature, and the traffic accident data includes visible light intensity, ground friction and accident-related monitoring video, and the accident-related monitoring video is represented by the monitoring video corresponding to the first time period before the time point of the accident occurrence, and the first time period is obtained through big data analysis.

3. A road traffic remote monitoring and early warning method as claimed in claim 1, characterized in that: The step S200 includes the following sub-steps: Step S201, setting visible light and ground friction as first classification conditions, setting a first value as a visible light threshold, and setting a second value as a ground friction threshold, performing a first classification on traffic accident information according to the first classification conditions to obtain first classification accidents, wherein the first classification accidents include first classification accidents in haze weather, first classification accidents in non-haze weather, first classification accidents in rainy and snowy weather, and first classification accidents in sunny weather; Step S202, performing frame segmentation and windowing processing on the accident-related surveillance video to obtain a surveillance image, and extracting a first key feature of the surveillance image, wherein the first key feature is a graphic feature, and the first key feature includes a person, a cat, a dog, a plastic bag, a leaf, and air; Step S203: analyzing the interference cause according to the first key feature, wherein the analysis logic of the interference cause includes: Obtain the first key feature. When the first key feature is air, set the interference cause to no interference. When the first key feature is a person, a cat, or a dog, set the interference cause to animal interference. When the first key feature is a plastic bag or a leaf, set the interference cause to still object interference. Step S204, setting the interference cause as a second classification condition, and performing a second classification on the first classification accident according to the second classification condition to obtain a second classification accident, wherein the second classification accident includes an animal interference second classification accident and a static interference second classification accident, and the second classification accident corresponds to the first classification accident Step S205, counting the total number of traffic accidents on the road section to be tested, recorded as the first number, counting the number of second-category accidents, recorded as the second number, calculating the ratio of the first number to the second number, recorded as the first accident probability, and performing a first combination of the second-category accidents and the first accident probability.

4. A road traffic remote monitoring and early warning method as claimed in claim 3, characterized in that: The classification logic of the first classification accident includes: Obtain visible light and ground friction, set a first value as a visible light threshold, and set a second value as a ground friction threshold, wherein the first value and the second value are obtained through big data analysis, and according to historical meteorological information, respectively calculate an average value of visible light corresponding to haze weather and an average value of ground friction corresponding to rainy and snowy weather, and set the average value of visible light corresponding to haze weather as the visible light threshold, and set the average value of ground friction corresponding to rainy and snowy weather as the ground friction threshold; Compare the visible light with the first value, and when the visible light is less than or equal to the first value, set the first category accident as a first category accident in haze weather; when the visible light is greater than the first value, set the first category accident as a first category accident in non-haze weather; The ground friction force is compared with the second value. When the ground friction force is greater than or equal to the second value, the first category accident is set as a first category accident in rainy and snowy weather. When the ground friction force is less than the second value, the first category accident is set as a first category accident in sunny weather. Among them, the first category accidents related to visible light and the first category accidents related to ground friction can be superimposed.

5. A road traffic remote monitoring and early warning method as claimed in claim 3, characterized in that: The extraction logic of the first key feature includes: Call the graphic database, select any picture in the graphic database, record it as the standard picture, extract the graphic feature of the standard picture, obtain the monitoring image, extract the graphic feature of the monitoring image, calculate the similarity between the graphic feature of the monitoring image and the graphic feature of the standard picture by the cosine similarity formula, traverse the similarity, sort the similarity in descending order, select the standard picture corresponding to the similarity with the largest value, set it as the first key image, set the graphic feature corresponding to the first key image as the first key feature, set the similarity threshold, compare the similarity with the similarity threshold, and when all similarities are less than the similarity threshold, set the first key feature to air.

6. A road traffic remote monitoring and early warning method as claimed in claim 1, characterized in that: The step S300 includes the following sub-steps: Step S301, obtain a monitoring image, obtain a time point corresponding to the monitoring image, record it as a key point, track the accident vehicle before the accident according to the accident vehicle license plate, match the position coordinates of the accident vehicle at the key point, the position coordinates are expressed as the coordinates of the center of gravity of the accident vehicle, set the position coordinates as the vehicle trajectory point, calculate the speed of the accident vehicle, and calculate the average value of the speed of the accident vehicle between the key points, set the average value of the speed of the accident vehicle between the key points as the average speed of the vehicle, set a horizontal road surface baseline, obtain the angle between the front of the accident vehicle and the horizontal road surface baseline, and calculate the average value of the angle between the key points, set the average value of the angle between the key points as the average steering angle of the vehicle, and draw a time-space distribution diagram of vehicle travel; The vehicle travel time-space distribution diagram includes vehicle track points, vehicle track curves, vehicle track direction arrows, vehicle average speed and vehicle average steering angle; Step S302, extracting a second key feature of the vehicle travel time-space distribution diagram, where the second key feature is a numerical feature, analyzing the cause of the accident according to the second key feature, setting the cause of the accident as a third classification condition, performing a third classification on the second classification accident, and obtaining a third classification accident; The third category accidents correspond to the second category accidents, and the third category accidents include overspeed third category accidents, oversteering third category accidents, normal speed third category accidents and normal steering category accidents. The steering-related third category accidents and speed-related third category accidents can be superimposed. Step S303, counting the number of accidents of the third category, recording it as the third number, calculating the ratio of the third number to the first number, recording it as the second accident probability, calculating the product of the second accident probability and the first accident probability, recording it as the comprehensive accident probability, and performing a second combination of the third accident category and the comprehensive accident probability; The vehicle trajectory points are obtained according to the position coordinates and the point drawing method. The vehicle trajectory curve is obtained by inserting n supplementary points between adjacent trajectory points through the interpolation method and connecting them through a smooth curve. The adjacent trajectory points are represented as position coordinate points corresponding to the key points in the time sequence. The smooth curve is represented as a function corresponding to the vehicle trajectory curve having a first-order continuous derivative; The extraction logic of the second key feature includes: Obtaining an average vehicle speed and an average vehicle steering angle, setting a third value as a vehicle average speed threshold, and setting a fourth value as a vehicle average steering angle threshold, wherein the third value and the fourth value are obtained through big data analysis; The average speed of the vehicle is compared with the third value, and when the average speed of the vehicle is greater than the third value, the second key feature is set to be too fast, otherwise, the second key feature is set to be normal; The average steering angle of the vehicle is compared with the fourth value. When the average steering angle of the vehicle is greater than the fourth value, the second feature is set to excessive steering; otherwise, the second key feature is set to normal steering.

7. A road traffic remote monitoring and early warning method as claimed in claim 6, characterized in that: The step S400 includes the following sub-steps: Step S401, establishing a second mapping relationship between the first key feature, the second key feature and the comprehensive accident probability; Step S402, obtaining the current surveillance video, extracting the current second key feature and the current first key feature according to the current surveillance video, performing a first match between the current first key feature and the first key feature, and performing a second match between the current second key feature and the second key feature, wherein the first match of the current first key feature is the same as the extraction logic of the first key feature, and the second match of the current second key feature is the same as the extraction logic of the second key feature; Step S403, obtaining a comprehensive accident probability corresponding to the result of the first match and the result of the second match, sending a first warning signal according to the comprehensive accident probability, and setting a fifth value and a sixth value as a probability threshold, wherein the fifth value and the sixth value are both distributed between 0 and 1, and the fifth value is less than the sixth value; When the comprehensive accident probability is less than or equal to the fifth value, a warning signal of the first brightness and the first color is issued, and the corresponding comprehensive accident probability is displayed; when the comprehensive accident probability is greater than the fifth value and less than or equal to the sixth value, a warning signal of the second brightness and the second color is issued, and the corresponding comprehensive accident probability is displayed; when the comprehensive accident probability is greater than the sixth value, a warning signal of the third brightness and the third color is issued, and the corresponding comprehensive accident probability is displayed, wherein the corresponding comprehensive accident probability is displayed through the OLED display screen.

8. A road traffic remote monitoring and early warning system, characterized in that: The system is used to execute the method in claim 1, comprising a collection module, an analysis module and an early warning module; The acquisition module obtains information of the road section to be tested and traffic accident data of the road section to be tested, wherein the traffic accident data includes visible light intensity, ground friction and accident-related monitoring video; The analysis module sets visible light intensity and ground friction as the first classification condition, performs a first classification on the traffic accident information according to the first classification condition to obtain a first classification accident, extracts a first key feature of the accident-related monitoring video, analyzes the interference cause according to the first key feature, sets the interference cause as the second classification condition, performs a second classification on the first classification accident according to the second classification condition to obtain a second classification accident, counts the total number of traffic accidents on the road section to be tested, recorded as the first number, counts the number of second classification accidents, recorded as the second number, calculates the ratio of the first number to the second number, recorded as the first accident probability, and performs a first combination of the second classification accident and the first accident probability, draws a vehicle travel time-space distribution map according to the accident-related monitoring video, extracts the second key feature of the vehicle travel time-space distribution map, analyzes the cause of the accident according to the second key feature, sets the cause of the accident as the third classification condition, performs a third classification on the second classification accident, obtains a third classification accident, counts the number of third classification accidents, recorded as the third number, calculates the ratio of the third number to the first number, recorded as the second accident probability, calculates the comprehensive accident probability according to the second accident probability, and performs a second combination of the third accident classification and the comprehensive accident probability; The early warning module establishes a second mapping relationship between the first key feature, the second key feature and the comprehensive accident probability, obtains the current comprehensive accident probability according to the first mapping relationship, and issues an early warning according to the comprehensive accident probability.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, a road traffic remote monitoring and early warning method as described in any one of claims 1-7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the road traffic remote monitoring and early warning method according to any one of claims 1 to 7 is implemented.

Images