A road traffic safety dynamic early warning system

Abstract

The present invention discloses a dynamic early warning system for road traffic safety, which relates to the technical field of traffic monitoring. When the system is in operation, the pedestrian behavior data X at the traffic intersection is collected in real time through the pedestrian behavior monitoring module, and combined with high-precision image acquisition equipment, the accurate judgment of pedestrian behavior is ensured. The risk behavior determination module compares the acquired behavior data X with the traffic light data, accurately identifies the potential risk state of pedestrians, and effectively reduces potential traffic safety hazards by triggering the early warning mechanism. With the triggering of the early warning mechanism, different levels of early warning prompt information are generated according to the early warning index Vw of the pedestrian passing by, providing timely warnings for traffic management departments and drivers, and further improving the responsiveness of traffic safety management. It can be adaptively optimized and adjusted according to the current traffic light timing at the traffic intersection, so that traffic fluency and safety are optimally balanced.

Description

Technical Field

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

[0002] Road traffic safety is an important area in urban management and traffic management, covering traffic signal control, traffic flow monitoring, accident warning, driver behavior monitoring and other aspects. With the acceleration of urbanization, traffic flow continues to increase, resulting in increasingly complex road traffic safety issues. Especially in busy urban intersections, crossroads or traffic-intensive areas such as schools and commercial areas, the interaction between pedestrians and vehicles becomes more complicated.

[0003] At present, the traffic management systems in most cities mainly rely on fixed timing of traffic lights, manual command of traffic police and traditional pedestrian crossing lights to ensure pedestrian safety. However, these measures have obvious limitations in some busy intersections or traffic-intensive areas. For example, the timing of traffic lights at many intersections is only adjusted according to the preset cycle, without dynamic adjustment in real time according to traffic flow, pedestrian behavior, and emergencies. This method often leads to conflicts between pedestrians and vehicles, especially during peak hours, when pedestrians do not obey traffic rules or rush to cross the road, they may collide with vehicles. Under traditional traffic signal control, the signal cycle is fixed and cannot respond to emergencies at intersections in a timely manner, posing a major safety hazard. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a road traffic safety dynamic warning system, which solves the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a road traffic safety dynamic warning system, including a pedestrian behavior monitoring module, a risk behavior determination module, a behavior warning module, a signal optimization module and a monitoring and evaluation module;

[0006] The pedestrian behavior monitoring module collects traffic intersection information and traffic light data through image acquisition equipment installed at the traffic intersection, including real-time collection of pedestrian behavior data X;

[0007] The risk behavior determination module compares the acquired behavior data X with the traffic light data to determine the risk status of the pedestrians passing by, and triggers the early warning mechanism according to the risk status result;

[0008] After the warning mechanism is triggered, the behavior warning module obtains the warning index Vw of the pedestrian passing by, and generates warning prompt information of different levels at the traffic intersection according to the warning index Vw;

[0009] After the early warning mechanism is triggered, the signal optimization module adaptively optimizes the timing information according to the current traffic light timing at the traffic intersection to obtain an optimized signal light timing data group;

[0010] After the early warning mechanism is triggered, the monitoring and evaluation module collects pedestrian flow and vehicle flow information at the traffic intersection at a fixed time period, and compares and evaluates the pedestrian flow and vehicle flow at the traffic intersection before the early warning mechanism is triggered, obtains the dynamic adjustment status results of the traffic intersection, and adaptively adjusts the early warning mechanism according to the dynamic adjustment status results of the traffic intersection.

[0011] Preferably, the pedestrian behavior monitoring module includes an image acquisition unit and a behavior recognition unit;

[0012] The image acquisition unit acquires traffic intersection information and traffic light data through an image acquisition device installed at the traffic intersection, including using a traffic monitoring camera and a traffic infrared sensor to acquire real-time video data of the traffic intersection, and performs frame processing on the real-time video data to acquire a set of continuous frames, which are marked as image data I;

[0013] Among them, the traffic light data includes the current traffic light status and the traffic light timing data;

[0014] The behavior recognition unit extracts pedestrian behavior features from the image data I to obtain the pedestrian feature vector Fh, and calculates the behavior risk value Rb based on the pedestrian feature vector and the traffic light state in the traffic light data. The behavior risk value Rb is used to quantify and measure whether there is any irregularity in the pedestrian's behavior. The behavior risk values ​​Rb of 1 to n pedestrians are integrated to generate behavior data X={Rb1, Rb2, ..., Rbn}.

[0015] Preferably, the pedestrian feature vector Fh is specifically obtained by detecting and locating pedestrians in the image data I using the YOLO target detection algorithm, and then calculating the speed and direction of the pedestrian using the optical flow method to generate the pedestrian feature vector Fh, including the pedestrian position Ph, the pedestrian speed Vh and the walking direction βh;

[0016] The behavior speed Vh is obtained by the following calculation formula:

[0017] ;

[0018] Where △Ph represents the displacement of the pedestrian in the time interval △t, △t represents the time interval, and the displacement △Ph is calculated by The calculation formula is obtained, where x(t) and y(t) represent the position coordinates of the pedestrian at time t, and x(t-1) and y(t-1) represent the position coordinates of the pedestrian at time t-1;

[0019] The walking direction βh is obtained by the following calculation formula:

[0020] ;

[0021] Where, arctan represents the inverse tangent function;

[0022] The behavior risk value Rb is obtained by the following calculation formula:

[0023] ;

[0024] Where r1, r2 and r3 represent weight coefficients, and r1, r2 and r3∈[0,1], r1+r2+r3=1, represents the pedestrian speed contribution, represents the pedestrian direction contribution, represents the traffic light state contribution item, Vmax represents the preset pedestrian speed upper limit of the traffic intersection, βsignal represents the direction of the traffic light at the traffic intersection, cos represents the cosine function, and Slight represents the traffic light state encoding value.

[0025] Preferably, the risk behavior determination module includes a trigger unit;

[0026] The trigger unit traverses the behavior risk value Rb(i) of the ith pedestrian among 1 to n pedestrians in the behavior data X, and compares it with the traffic light data and the warning trigger threshold value Yz, wherein the traffic light data includes the traffic light state code value Slight, determines the risk state result of the behavior risk value Rb(i) of the ith pedestrian passing through the traffic intersection, and triggers the warning mechanism according to the risk state result, obtains the total number of people Cz that triggers the warning mechanism by aggregating the risk state results of 1 to n pedestrians passing through the traffic intersection, and compares it with the preset traffic intersection crowd risk threshold RQ to obtain the crowd risk state result of the traffic intersection;

[0027] The total number of people Cz is The calculation formula is obtained, C(i) represents the trigger factor of the i-th pedestrian;

[0028] The population risk status results are obtained by the following comparison method:

[0029] When the total number of people Cz ≥ the crowd risk threshold RQ at the traffic intersection, the crowd risk status result is obtained as the risk result, triggering the early warning mechanism;

[0030] When the total number of people Cz is less than the crowd risk threshold RQ at the traffic intersection, the crowd risk status result is obtained as a risk-free result.

[0031] Preferably, the risk status result is obtained by the following comparison method:

[0032] When the behavior risk value Rb(i) ≥ the warning trigger threshold Yz, and the traffic light state code value Slight = 1, the risk state result of the i-th pedestrian is obtained as a risk state, the warning mechanism is triggered, and the trigger factor C = 3 of the i-th pedestrian is marked;

[0033] When the behavior risk value Rb(i) ≥ the warning trigger threshold Yz, and the traffic light state code value Slight = 0, the risk state result of the i-th pedestrian is obtained as no risk state, the warning mechanism is not triggered, and the trigger factor C of the i-th pedestrian is marked as 1;

[0034] When the behavior risk value Rb(i) is less than the warning trigger threshold Yz, and the traffic light state code value Slight=1, the risk state result of the i-th pedestrian is obtained as no risk state, the warning mechanism is not triggered, and the trigger factor C of the i-th pedestrian is marked as 0;

[0035] When the behavior risk value Rb(i) is less than the warning trigger threshold Yz, and the traffic light state code value Slight=0, the risk state result of the i-th pedestrian is obtained as no risk state, the warning mechanism is not triggered, and the trigger factor C=0 of the i-th pedestrian is marked.

[0036] Preferably, the behavior warning module includes a warning generation unit;

[0037] After the warning mechanism is triggered, the warning generation unit obtains the warning index Vw of the i-th pedestrian passing through the traffic intersection by fitting the behavior risk value Rb(i) of the i-th pedestrian and the trigger factor C(i) of the i-th pedestrian, and generates warning prompt information of different levels for the traffic intersection according to the comparison between the warning index Vw and the preset traffic intersection warning generation threshold YZ;

[0038] The warning index Vw is The calculation formula is obtained, Max represents the maximum value function, Max (C*Rb) represents the upper limit trigger factor C and the behavior risk value Rb in the behavior data X, which is used to standardize the warning index Vw to the range of [0, 1].

[0039] Preferably, the warning prompt information is obtained by the following generation method:

[0040] When the warning index Vw is less than the traffic intersection warning generation threshold YZ, the first-level warning prompt information is obtained, and the relevant prompt slogans of civilized traffic and courtesy to pedestrians are displayed;

[0041] When the warning index Vw ≥ the traffic intersection warning generation threshold YZ, obtain the secondary warning prompt information, marginally display the relevant prompt slogans of civilized traffic and courtesy to pedestrians, and at the same time remind pedestrians to pay attention to traffic lights and drivers to pay attention to pedestrians;

[0042] When the warning index Vw ≥ twice the traffic intersection warning generation threshold YZ, the third-level warning prompt information is obtained, and the relevant prompt slogans of civilized traffic and courtesy to pedestrians are displayed on the edge. At the same time, pedestrians are reminded to pay attention to traffic lights and drivers are reminded to pay attention to pedestrians, and the timing of signal lights is adjusted.

[0043] Preferably, the signal optimization module includes a signal light timing optimization unit;

[0044] After the early warning mechanism is triggered, the traffic light timing optimization unit performs adaptive optimization of the timing information according to the traffic light timing at the current traffic intersection, wherein the traffic light timing includes the green light time Tg, the yellow light time Ty and the red light time Tr, and performs fitting calculation with the dynamic adjustment ratio Pa to obtain an optimized traffic light timing data group, including the optimized green light time YTg, the optimized yellow light time YTy and the optimized red light time YTr, and generates a control instruction for the optimized traffic light timing data group to control the traffic light timing at the current traffic intersection;

[0045] The dynamic adjustment ratio Pa is obtained by the following calculation formula:

[0046] ;

[0047] In the formula, γ represents the adjustment coefficient, which is specifically used to adjust the adjustment range of the signal light timing;

[0048] The optimized green light time YTg is obtained by the calculation formula YTg=Tg*(1+Pa);

[0049] The optimized yellow light time YTy is obtained by the calculation formula YTy=Ty*(1+Pa);

[0050] The optimized red light time YTr is obtained by the calculation formula YTr=Tr*(1+Pa).

[0051] Preferably, the monitoring and evaluation module includes a flow collection unit and a dynamic adjustment unit;

[0052] After the early warning mechanism is triggered, the flow collection unit collects pedestrian flow and vehicle flow information at the traffic intersection at a fixed time period, including pedestrian flow Pflow and vehicle flow Vflow;

[0053] The pedestrian flow Pflow is obtained by the following calculation formula:

[0054] ;

[0055] Where Pflow(t) represents the pedestrian flow at time t, Pcount(i, t) represents the count of the i-th pedestrian at time t, and m represents the total number of pedestrians passing through the traffic intersection at time t;

[0056] The vehicle flow Vflow is obtained by the following calculation formula:

[0057] ;

[0058] Where Vflow(t) represents the vehicle flow at time t, Vcount(j, t) represents the count of the jth vehicle at time t, and w represents the total number of vehicles passing the traffic intersection at time t.

[0059] Preferably, the dynamic adjustment unit compares and evaluates the pedestrian flow Pflow and the vehicle flow Vflow with the pedestrian flow and the vehicle flow at the traffic intersection before the early warning mechanism is triggered, wherein the pedestrian flow and the vehicle flow at the traffic intersection before the early warning mechanism is triggered are marked as the front pedestrian flow QPflow and the front vehicle flow QVflow, and obtains the difference between the pedestrian flow Pflow and the vehicle flow Vflow and the front pedestrian flow QPflow and the front vehicle flow QVflow, which are marked as the pedestrian flow change △Pflow and the vehicle flow change △Vflow;

[0060] Then, the dynamic adjustment coefficient Rad is obtained by calculating the pedestrian flow change △Pflow and the vehicle flow change △Vflow with the pedestrian flow Pflow and the vehicle flow Vflow, and the dynamic adjustment state result of the traffic intersection is obtained, and the early warning mechanism is adaptively adjusted according to the dynamic adjustment state result of the traffic intersection;

[0061] The dynamic adjustment coefficient Rad is obtained by Obtain calculation formula;

[0062] The dynamic adjustment state result of the traffic intersection is obtained by the following comparison method:

[0063] When the dynamic adjustment coefficient Rad>0.1, it means that the traffic flow is increasing, and the signal light timing is adjusted proportionally, including optimizing the green light time YTg, optimizing the yellow light time YTy and optimizing the red light time YTr;

[0064] When the dynamic adjustment coefficient Rad is less than -0.1, it indicates that the flow rate is decreasing, and the traffic light timing of the initial traffic intersection includes the green light time Tg, the yellow light time Ty and the red light time Tr.

[0065] The present invention provides a road traffic safety dynamic warning system, which has the following beneficial effects:

[0066] (1) When the system is running, the pedestrian behavior monitoring module collects pedestrian behavior data X at the traffic intersection in real time, and combines it with high-precision image acquisition equipment to ensure accurate judgment of pedestrian behavior. The risk behavior determination module compares the acquired behavior data X with the traffic light data to accurately identify the potential risk status of pedestrians, and effectively reduces potential traffic safety hazards by triggering the early warning mechanism. With the triggering of the early warning mechanism, the behavior early warning module will generate different levels of early warning information according to the early warning index Vw of the pedestrian passing by, providing timely warnings for traffic management departments and drivers, further improving the responsiveness of traffic safety management. At the same time, the signal optimization module can adaptively optimize and adjust the current traffic light timing at the traffic intersection, so that traffic smoothness and safety are optimally balanced, avoiding traffic congestion and safety hazards caused by unreasonable signal timing.

[0067] (2) By real-time evaluation and identification of potential risk behaviors of pedestrians at traffic intersections, and triggering an accurate early warning mechanism. The trigger unit traverses all pedestrians in the behavior data X, and compares each pedestrian's behavior risk value Rb(i) with the traffic light state code value Slight and the early warning trigger threshold Yz to comprehensively determine whether the pedestrian is in a dangerous state. The system will trigger the early warning mechanism, and timely adjust the traffic light timing and issue relevant prompt information to minimize traffic safety risks. The advantage lies in its highly intelligent behavior analysis and dynamic response capabilities. By accurately calculating and evaluating the risk behavior status of each pedestrian. On this basis, the system can generate targeted traffic prompt information according to different early warning levels, which not only improves the accuracy of pedestrian behavior monitoring, but also optimizes the real-time nature of traffic light timing and pedestrian safety management.

[0068] (3) Dynamically adjust the timing of traffic lights according to different risk levels and flow conditions, thereby improving traffic flow and safety. After the warning mechanism is triggered, the system first determines the risk level of the current traffic intersection through the warning index Vw, and generates different levels of warning prompt information according to the set warning generation threshold YZ. By periodically collecting and evaluating the changes in pedestrian flow Pflow and vehicle flow Vflow, the timing of traffic lights is dynamically adjusted. By comparing the flow changes before and after the warning mechanism is triggered, the dynamic adjustment coefficient Rad is calculated, and then it is decided whether to further optimize the timing of the traffic lights. This adjustment mechanism effectively improves the adaptability of traffic intersections, enabling them to automatically adjust the timing of traffic lights under different traffic flows and risk conditions, thereby improving traffic efficiency and pedestrian safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 The present invention is a schematic diagram of a block diagram of a dynamic early warning system for road traffic safety. DETAILED DESCRIPTION

[0070] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0071] Example 1

[0072] The present invention provides a road traffic safety dynamic warning system, please refer to Figure 1 , including pedestrian behavior monitoring module, risk behavior determination module, behavior warning module, signal optimization module and monitoring and evaluation module;

[0073] The pedestrian behavior monitoring module collects traffic intersection information and traffic light data through image acquisition equipment installed at the traffic intersection, including real-time collection of pedestrian behavior data X;

[0074] The risk behavior determination module compares the acquired behavior data X with the traffic light data to determine the risk status of the pedestrians passing by, and triggers the early warning mechanism according to the risk status result;

[0075] After the warning mechanism is triggered, the behavior warning module obtains the warning index Vw of the pedestrian passing by, and generates warning prompt information of different levels at the traffic intersection according to the warning index Vw;

[0076] After the early warning mechanism is triggered, the signal optimization module adaptively optimizes the timing information according to the current traffic light timing at the traffic intersection to obtain an optimized signal light timing data group;

[0077] After the early warning mechanism is triggered, the monitoring and evaluation module collects pedestrian flow and vehicle flow information at the traffic intersection at a fixed time period, and compares and evaluates the pedestrian flow and vehicle flow at the traffic intersection before the early warning mechanism is triggered, obtains the dynamic adjustment status results of the traffic intersection, and adaptively adjusts the early warning mechanism according to the dynamic adjustment status results of the traffic intersection.

[0078] In this embodiment, the pedestrian behavior monitoring module collects the pedestrian behavior data X at the traffic intersection in real time, and combines with high-precision image acquisition equipment to ensure accurate judgment of pedestrian behavior. The risk behavior determination module compares the obtained behavior data X with the traffic light data, accurately identifies the potential risk status of pedestrians, and effectively reduces potential traffic safety hazards by triggering the early warning mechanism. With the triggering of the early warning mechanism, the behavior early warning module will generate early warning prompt information of different levels according to the early warning index Vw of the pedestrian passing by, providing timely warnings for the traffic management department and the driver, and further improving the responsiveness of traffic safety management. At the same time, the signal optimization module can adaptively optimize and adjust the current traffic light timing at the traffic intersection, so that traffic smoothness and safety are optimally balanced, avoiding traffic congestion and safety hazards caused by unreasonable signal timing. The monitoring and evaluation module collects pedestrian and vehicle flow data at a fixed time period, compares and analyzes the flow before and after the early warning mechanism is triggered, evaluates and timely adjusts the traffic light timing and early warning mechanism, thereby ensuring the dynamic adaptability and sustainability of the system, especially in terms of dynamic evaluation and adaptive adjustment, it can flexibly adjust traffic signals and management strategies according to real-time data, avoiding the problems of being too rigid and lagging. Through real-time comprehensive analysis of pedestrian behavior, vehicle flow, and traffic light status, the system can accurately identify potential risks, issue warnings in a timely manner, and take measures to ensure the safety, smoothness, and efficiency of traffic. In addition, by comprehensively considering the pedestrian flow, vehicle flow, and signal light timing optimization at traffic intersections, the system avoids blind adjustments in traditional traffic signal control and improves the accuracy and intelligence of traffic management. This comprehensive and dynamic adjustment mechanism significantly improves traffic safety and the efficiency of intersection management, especially in high-risk periods and high-flow conditions, providing more reliable decision-making support for traffic management.

[0079] Example 2

[0080] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the pedestrian behavior monitoring module includes an image acquisition unit and a behavior recognition unit;

[0081] The image acquisition unit acquires traffic intersection information and traffic light data through an image acquisition device installed at the traffic intersection, including using a traffic monitoring camera and a traffic infrared sensor to acquire real-time video data of the traffic intersection, and performs frame processing on the real-time video data to acquire a set of continuous frames, which are marked as image data I;

[0082] Among them, the traffic light data includes the current traffic light status and the traffic light timing data;

[0083] The behavior recognition unit extracts pedestrian behavior features from the image data I to obtain the pedestrian feature vector Fh, and calculates the behavior risk value Rb based on the pedestrian feature vector and the traffic light state in the traffic light data. The behavior risk value Rb is used to quantify and measure whether there is any irregularity in the pedestrian's behavior. The behavior risk values ​​Rb of 1 to n pedestrians are integrated to generate behavior data X={Rb1, Rb2, ..., Rbn}.

[0084] The pedestrian feature vector Fh is specifically obtained by detecting and locating the pedestrian in the image data I using the YOLO target detection algorithm, and then calculating the pedestrian's speed and direction using the optical flow method to generate the pedestrian feature vector Fh, including the pedestrian's position Ph, pedestrian speed Vh and walking direction βh;

[0085] The behavior speed Vh is obtained by the following calculation formula:

[0086] ;

[0087] Where △Ph represents the displacement of the pedestrian in the time interval △t, △t represents the time interval, and the displacement △Ph is calculated by The calculation formula is obtained, where x(t) and y(t) represent the position coordinates of the pedestrian at time t, and x(t-1) and y(t-1) represent the position coordinates of the pedestrian at time t-1;

[0088] The walking direction βh is obtained by the following calculation formula:

[0089] ;

[0090] Wherein, arctan represents the inverse tangent function, which is specifically used to calculate the angle corresponding to the slope between the position at time t and the position at time t-1;

[0091] The behavior risk value Rb is obtained by the following calculation formula:

[0092] ;

[0093] Where r1, r2 and r3 represent weight coefficients, and r1, r2 and r3∈[0,1], r1+r2+r3=1, It represents the pedestrian speed contribution item, which is used to evaluate whether the pedestrian is speeding. At the same time, r1 is used to adjust the impact of speed on the behavior risk value Rb. It represents the pedestrian direction contribution item. The cosine value of the direction deviation is used to calculate the degree of deviation between the pedestrian's walking direction and the direction of the traffic light. At the same time, r2 is used to adjust the influence of the direction on the behavior risk value Rb. It represents the traffic light status contribution item, which is used to evaluate the risk contribution of pedestrians under different traffic lights. At the same time, r3 is used to adjust the impact of the traffic light status on the behavior risk value Rb. Vmax represents the preset upper limit of pedestrian speed at the traffic intersection. βsignal represents the direction of the traffic light at the traffic intersection, including the directions parallel and perpendicular to the road. It is used to calculate the degree of deviation between the pedestrian movement direction and the direction of the traffic light. cos represents the cosine function. Slight represents the traffic light status coding value, which is specifically reflected by the preset green light coding value of 0 and the red light coding value of 1. In the red light state, the risk assessment value will directly increase, thereby giving priority to triggering the early warning mechanism.

[0094] In this embodiment, the pedestrian information at the traffic intersection is captured in real time by the image acquisition unit, and the pedestrian behavior characteristics are extracted using the advanced YOLO target detection algorithm and optical flow method. Through this series of precise technical means, not only can the position, speed and direction of the pedestrian passing by be accurately identified, but also the pedestrian behavior risk value Rb can be calculated in real time to quantify whether its behavior complies with the specification. This behavior risk assessment mechanism based on real-time data analysis can dynamically monitor multiple pedestrians and promptly discover potential traffic safety risks. By integrating the behavior risk values ​​Rb of multiple pedestrians and generating behavior data X, the system can comprehensively assess the safety status of the intersection and provide data support for subsequent risk determination and early warning mechanisms. By considering the relationship between the pedestrian's movement speed, walking direction and the state of the signal light, the system can more comprehensively and meticulously assess whether the pedestrian's behavior is potentially dangerous. For example, when the pedestrian's behavior speed is too fast or inconsistent with the direction of the signal light, the system will significantly increase the risk assessment value of the pedestrian, thereby providing more accurate and real-time early warning information for traffic management.

[0095] Example 3

[0096] This embodiment is explained in Example 2. Please refer to Figure 1 ,Specifically: the risk behavior determination module includes a trigger unit;

[0097] The trigger unit traverses the behavior risk value Rb(i) of the ith pedestrian among 1 to n pedestrians in the behavior data X, and compares it with the traffic light data and the warning trigger threshold value Yz, wherein the traffic light data includes the traffic light state code value Slight, determines the risk state result of the behavior risk value Rb(i) of the ith pedestrian passing through the traffic intersection, and triggers the warning mechanism according to the risk state result, obtains the total number of people Cz that triggers the warning mechanism by aggregating the risk state results of 1 to n pedestrians passing through the traffic intersection, and compares it with the preset traffic intersection crowd risk threshold RQ to obtain the crowd risk state result of the traffic intersection;

[0098] The total number of people Cz is The calculation formula is obtained, C(i) represents the trigger factor of the i-th pedestrian;

[0099] The population risk status results are obtained by the following comparison method:

[0100] When the total number of people Cz ≥ the crowd risk threshold RQ at the traffic intersection, the crowd risk status result is obtained as the risk result, triggering the early warning mechanism, indicating that there are multiple pedestrians in a dangerous behavior state, adjusting the traffic signal timing and issuing reminder information;

[0101] When the total number of people Cz is less than the crowd risk threshold RQ at the traffic intersection, the crowd risk status result is obtained as a risk-free result.

[0102] The risk status results are obtained by the following comparison method:

[0103] When the behavior risk value Rb(i) ≥ the warning trigger threshold Yz, and the traffic light state code value Slight=1, the risk state result of the i-th pedestrian is obtained as a risk state, triggering the warning mechanism, marking the trigger factor C=3 of the i-th pedestrian, reflecting that the i-th pedestrian is in a high-risk behavior under the red light state, indicating that the pedestrian's behavior may cause danger, and it is necessary to remind the driver or control the traffic light in time. At the same time, the pedestrian feature vector Fh of the i-th pedestrian is extracted to remind pedestrians and drivers on the traffic intersection display;

[0104] When the behavior risk value Rb(i) ≥ the warning trigger threshold Yz, and the traffic light state code value Slight=0, the risk state result of the i-th pedestrian is obtained as no risk state, and the warning mechanism is not triggered. The trigger factor C=1 of the i-th pedestrian is marked, reflecting that the i-th pedestrian does not behave properly under the green light state, and the pedestrian's behavior may cause danger, and the driver needs to be reminded in time. At the same time, the pedestrian feature vector Fh of the i-th pedestrian is extracted to prompt pedestrians and drivers on the traffic intersection display;

[0105] When the behavior risk value Rb(i) is less than the warning trigger threshold Yz, and the traffic light state code value Slight=1, the risk state result of the i-th pedestrian is obtained as no risk state, and the warning mechanism is not triggered. The trigger factor C=0 of the i-th pedestrian is marked, and the pedestrian's behavior is not illegal, or the behavior risk is low;

[0106] When the behavior risk value Rb(i) is less than the warning trigger threshold Yz, and the traffic light status code value Slight=0, the risk status result of the i-th pedestrian is obtained as no risk status, the warning mechanism is not triggered, and the trigger factor C=0 of the i-th pedestrian is marked, indicating that the pedestrian has not violated the rules and has no risk.

[0107] The behavior warning module includes a warning generation unit;

[0108] After the warning mechanism is triggered, the warning generation unit obtains the warning index Vw of the i-th pedestrian passing through the traffic intersection by fitting the behavior risk value Rb(i) of the i-th pedestrian and the trigger factor C(i) of the i-th pedestrian, and generates warning prompt information of different levels for the traffic intersection according to the comparison between the warning index Vw and the preset traffic intersection warning generation threshold YZ;

[0109] The warning index Vw is The calculation formula is obtained, Max represents the maximum value function, Max (C*Rb) represents the upper limit trigger factor C and the behavior risk value Rb in the behavior data X, which is used to standardize the warning index Vw to the range of [0, 1].

[0110] In this embodiment, the potential risk behaviors of pedestrians at traffic intersections are evaluated and identified in real time, and an accurate early warning mechanism is triggered. The trigger unit traverses all pedestrians in the behavior data X, and compares the behavior risk value Rb (i) of each pedestrian with the traffic light state code value Slight and the early warning trigger threshold Yz to comprehensively determine whether the pedestrian is in a dangerous state. In addition, the system can also generate a total number of people Cz by aggregating the risk status results of multiple pedestrians, and compare it with the preset traffic intersection crowd risk threshold RQ to determine whether there is an overall risk at the traffic intersection. When the total number of people Cz reaches the preset threshold, the system will trigger the early warning mechanism, and timely adjust the traffic light timing and issue relevant prompt information to minimize traffic safety risks. The advantage lies in its highly intelligent behavior analysis and dynamic response capabilities. By accurately calculating and evaluating the risk behavior status of each pedestrian, the system can assign a corresponding trigger factor C (i) to each pedestrian under different traffic light states, and generate an early warning index Vw in combination with the pedestrian behavior risk value Rb (i). This warning index Vw is further standardized and adjusted to the range of [0, 1], making the warning mechanism more flexible and with higher response accuracy. On this basis, the system can generate targeted traffic prompt information according to different warning levels, which not only improves the accuracy of pedestrian behavior monitoring, but also optimizes the real-time timing of traffic lights and pedestrian safety management. Through this intelligent warning and dynamic adjustment mechanism, the risk management ability of traffic intersections has been significantly improved, effectively reducing the risk of traffic accidents, while enhancing the flexibility and intelligence level of the transportation system.

[0111] Example 4

[0112] This embodiment is explained in Example 3, please refer to Figure 1 Specifically: the warning prompt information is obtained by the following generation method:

[0113] When the warning index Vw is less than the traffic intersection warning generation threshold YZ, the first-level warning prompt information is obtained, and the relevant prompt slogans of civilized traffic and courtesy to pedestrians are displayed;

[0114] When the warning index Vw ≥ the traffic intersection warning generation threshold YZ, obtain the secondary warning prompt information, marginally display the relevant prompt slogans of civilized traffic and courtesy to pedestrians, and at the same time remind pedestrians to pay attention to traffic lights and drivers to pay attention to pedestrians;

[0115] When the warning index Vw ≥ twice the traffic intersection warning generation threshold YZ, the third-level warning prompt information is obtained, and the relevant prompt slogans of civilized traffic and courtesy to pedestrians are displayed on the edge. At the same time, pedestrians are reminded to pay attention to traffic lights and drivers are reminded to pay attention to pedestrians, and the timing of signal lights is adjusted.

[0116] The signal optimization module includes a signal light timing optimization unit;

[0117] After the early warning mechanism is triggered, the traffic light timing optimization unit performs adaptive optimization of the timing information according to the traffic light timing at the current traffic intersection, wherein the traffic light timing includes the green light time Tg, the yellow light time Ty and the red light time Tr, and performs fitting calculation with the dynamic adjustment ratio Pa to obtain an optimized traffic light timing data group, including the optimized green light time YTg, the optimized yellow light time YTy and the optimized red light time YTr, and generates a control instruction for the optimized traffic light timing data group to control the traffic light timing at the current traffic intersection;

[0118] The dynamic adjustment ratio Pa is obtained by the following calculation formula:

[0119] ;

[0120] In the formula, γ represents the adjustment coefficient, which is specifically used to adjust the adjustment range of the signal light timing;

[0121] The optimized green light time YTg is obtained by the calculation formula YTg=Tg*(1+Pa);

[0122] The optimized yellow light time YTy is obtained by the calculation formula YTy=Ty*(1+Pa);

[0123] The optimized red light time YTr is obtained by the calculation formula YTr=Tr*(1+Pa).

[0124] The monitoring and evaluation module includes a flow collection unit and a dynamic adjustment unit;

[0125] After the early warning mechanism is triggered, the flow collection unit collects pedestrian flow and vehicle flow information at the traffic intersection at a fixed time period, including pedestrian flow Pflow and vehicle flow Vflow;

[0126] The pedestrian flow Pflow is obtained by the following calculation formula:

[0127] ;

[0128] Where Pflow(t) represents the pedestrian flow at time t, Pcount(i, t) represents the count of the i-th pedestrian at time t, and m represents the total number of pedestrians passing through the traffic intersection at time t;

[0129] The vehicle flow Vflow is obtained by the following calculation formula:

[0130] ;

[0131] Where Vflow(t) represents the vehicle flow at time t, Vcount(j, t) represents the count of the jth vehicle at time t, and w represents the total number of vehicles passing the traffic intersection at time t.

[0132] The dynamic adjustment unit compares and evaluates the pedestrian flow Pflow and the vehicle flow Vflow with the pedestrian flow and the vehicle flow at the traffic intersection before the early warning mechanism is triggered, wherein the pedestrian flow and the vehicle flow at the traffic intersection before the early warning mechanism is triggered are marked as the previous pedestrian flow QPflow and the previous vehicle flow QVflow, and obtains the difference between the pedestrian flow Pflow and the vehicle flow Vflow and the previous pedestrian flow QPflow and the previous vehicle flow QVflow, which are marked as the pedestrian flow change △Pflow and the vehicle flow change △Vflow;

[0133] Then, the dynamic adjustment coefficient Rad is obtained by calculating the pedestrian flow change △Pflow and the vehicle flow change △Vflow with the pedestrian flow Pflow and the vehicle flow Vflow, and the dynamic adjustment state result of the traffic intersection is obtained, and the early warning mechanism is adaptively adjusted according to the dynamic adjustment state result of the traffic intersection;

[0134] The dynamic adjustment coefficient Rad is obtained by Obtain calculation formula;

[0135] The dynamic adjustment state result of the traffic intersection is obtained by the following comparison method:

[0136] When the dynamic adjustment coefficient Rad>0.1, it means that the traffic flow is increasing, and the signal light timing is adjusted proportionally, including optimizing the green light time YTg, optimizing the yellow light time YTy and optimizing the red light time YTr;

[0137] When the dynamic adjustment coefficient Rad is less than -0.1, it indicates that the flow rate is decreasing, and the traffic light timing of the initial traffic intersection includes the green light time Tg, the yellow light time Ty and the red light time Tr.

[0138] In this embodiment, the traffic light timing is dynamically adjusted according to different risk levels and traffic conditions, thereby improving traffic fluency and safety. After the early warning mechanism is triggered, the system first determines the risk level of the current traffic intersection through the early warning index Vw, and generates different levels of early warning prompt information according to the set early warning generation threshold YZ. When the Vw value is lower than YZ, the system only provides basic civilized traffic prompts; when the Vw value reaches or exceeds YZ, the system will further prompt pedestrians to pay attention to traffic lights and remind drivers to pay attention to pedestrians. Finally, when the Vw value exceeds twice YZ, the system will not only increase warning information, but also adjust the timing of the traffic lights to ensure the safety of pedestrians in high-risk situations. On this basis, the green, yellow and red light durations of the traffic lights are optimized by dynamically adjusting the ratio Pa to adapt to real-time traffic flow changes. The optimized green, yellow and red light times are calculated by formulas, respectively, to ensure that the traffic light timing can be accurately adjusted according to traffic flow and pedestrian risk index. When traffic flow shows an upward trend, the optimized signal light timing can effectively extend the green light time and reduce waiting time; when the flow decreases, the system will restore the standard signal light timing to avoid unnecessary waste of resources. In addition, by periodically collecting and evaluating the changes in pedestrian flow Pflow and vehicle flow Vflow, the traffic light timing is dynamically adjusted. By comparing the flow changes before and after the early warning mechanism is triggered, the dynamic adjustment coefficient Rad is calculated to determine whether to further optimize the signal light timing. This adjustment mechanism effectively improves the adaptability of traffic intersections, enabling them to automatically adjust the signal light timing under different traffic flow and risk conditions, improve traffic efficiency and pedestrian safety, and thus achieve the ability to flexibly respond to traffic conditions under different flow conditions, and can also accurately control the signal light timing, thereby reducing the probability of traffic accidents and improving the road's traffic efficiency.

[0139] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A road traffic safety dynamic warning system, characterized in that: It includes pedestrian behavior monitoring module, risk behavior determination module, behavior warning module, signal optimization module and monitoring and evaluation module; The pedestrian behavior monitoring module collects traffic intersection information and traffic light data through image acquisition equipment installed at the traffic intersection, including real-time collection of pedestrian behavior data X; The risk behavior determination module compares the acquired behavior data X with the traffic light data to determine the risk status of the pedestrians passing by, and triggers the early warning mechanism according to the risk status result; The risk behavior determination module includes a trigger unit; The trigger unit traverses the behavior risk value Rb(i) of the ith pedestrian among 1 to n pedestrians in the behavior data X, and compares it with the traffic light data and the warning trigger threshold value Yz, wherein the traffic light data includes the traffic light state code value Slight, determines the risk state result of the behavior risk value Rb(i) of the ith pedestrian passing through the traffic intersection, and triggers the warning mechanism according to the risk state result, obtains the total number of people Cz that triggers the warning mechanism by aggregating the risk state results of 1 to n pedestrians passing through the traffic intersection, and compares it with the preset traffic intersection crowd risk threshold RQ to obtain the crowd risk state result of the traffic intersection; The total number of people Cz is The calculation formula is obtained, C(i) represents the trigger factor of the i-th pedestrian; The population risk status results are obtained by the following comparison method: When the total number of people Cz ≥ the crowd risk threshold RQ at the traffic intersection, the crowd risk status result is obtained as the risk result, triggering the early warning mechanism; When the total number of people Cz is less than the crowd risk threshold RQ at the traffic intersection, the crowd risk status result is obtained as a risk-free result; The risk status results are obtained by the following comparison method: When the behavior risk value Rb(i) ≥ the warning trigger threshold Yz, and the traffic light state code value Slight = 1, the risk state result of the i-th pedestrian is obtained as a risk state, the warning mechanism is triggered, and the trigger factor C = 3 of the i-th pedestrian is marked; When the behavior risk value Rb(i) ≥ the warning trigger threshold Yz, and the traffic light state code value Slight = 0, the risk state result of the i-th pedestrian is obtained as no risk state, the warning mechanism is not triggered, and the trigger factor C of the i-th pedestrian is marked as 1; When the behavior risk value Rb(i) is less than the warning trigger threshold Yz, and the traffic light state code value Slight=1, the risk state result of the i-th pedestrian is obtained as no risk state, the warning mechanism is not triggered, and the trigger factor C of the i-th pedestrian is marked as 0; When the behavior risk value Rb(i) is less than the warning trigger threshold Yz, and the traffic light state code value Slight=0, the risk state result of the i-th pedestrian is obtained as no risk state, the warning mechanism is not triggered, and the trigger factor C of the i-th pedestrian is marked as 0; After the warning mechanism is triggered, the behavior warning module obtains the warning index Vw of the pedestrian passing by, and generates warning prompt information of different levels at the traffic intersection according to the warning index Vw; After the early warning mechanism is triggered, the signal optimization module adaptively optimizes the timing information according to the current traffic light timing at the traffic intersection to obtain an optimized signal light timing data group; After the early warning mechanism is triggered, the monitoring and evaluation module collects pedestrian flow and vehicle flow information at the traffic intersection at a fixed time period, and compares and evaluates the pedestrian flow and vehicle flow at the traffic intersection before the early warning mechanism is triggered, obtains the dynamic adjustment status results of the traffic intersection, and adaptively adjusts the early warning mechanism according to the dynamic adjustment status results of the traffic intersection.

2. A road traffic safety dynamic warning system according to claim 1, characterized in that: The pedestrian behavior monitoring module includes an image acquisition unit and a behavior recognition unit; The image acquisition unit acquires traffic intersection information and traffic light data through an image acquisition device installed at the traffic intersection, including using a traffic monitoring camera and a traffic infrared sensor to acquire real-time video data of the traffic intersection, and performs frame processing on the real-time video data to acquire a set of continuous frames, which are marked as image data I; Among them, the traffic light data includes the current traffic light status and the traffic light timing data; The behavior recognition unit extracts pedestrian behavior features from the image data I to obtain the pedestrian feature vector Fh, and calculates the behavior risk value Rb based on the pedestrian feature vector and the traffic light state in the traffic light data. The behavior risk value Rb is used to quantify and measure whether there is any irregularity in the pedestrian's behavior. The behavior risk values ​​Rb of 1 to n pedestrians are integrated to generate behavior data X={Rb1, Rb2, ..., Rbn}.

3. A road traffic safety dynamic early warning system according to claim 2, characterized in that: in, The pedestrian feature vector Fh is specifically obtained by detecting and locating pedestrians in the image data I using the YOLO target detection algorithm, and then calculating the pedestrian's speed and direction using the optical flow method to generate the pedestrian feature vector Fh, including the pedestrian's position Ph, pedestrian speed Vh and walking direction βh; The behavior speed Vh is obtained by the following calculation formula: ； Where △Ph represents the displacement of the pedestrian in the time interval △t, △t represents the time interval, and the displacement △Ph is calculated by The calculation formula is obtained, where x(t) and y(t) represent the position coordinates of the pedestrian at time t, and x(t-1) and y(t-1) represent the position coordinates of the pedestrian at time t-1; The walking direction βh is obtained by the following calculation formula: ； Where, arctan represents the inverse tangent function; The behavior risk value Rb is obtained by the following calculation formula: ； Where r1, r2 and r3 represent weight coefficients, and r1, r2 and r3∈[0,1], r1+r2+r3=1, represents the pedestrian speed contribution, represents the pedestrian direction contribution, represents the traffic light state contribution item, Vmax represents the preset pedestrian speed upper limit of the traffic intersection, βsignal represents the direction of the traffic light at the traffic intersection, cos represents the cosine function, and Slight represents the traffic light state encoding value.

4. A road traffic safety dynamic warning system according to claim 1, characterized in that: The behavior warning module includes a warning generation unit; After the warning mechanism is triggered, the warning generation unit obtains the warning index Vw of the i-th pedestrian passing through the traffic intersection by fitting the behavior risk value Rb(i) of the i-th pedestrian and the trigger factor C(i) of the i-th pedestrian, and generates warning prompt information of different levels for the traffic intersection according to the comparison between the warning index Vw and the preset traffic intersection warning generation threshold YZ; The warning index Vw is The calculation formula is obtained, Max represents the maximum value function, Max (C*Rb) represents the upper limit trigger factor C and the behavior risk value Rb in the behavior data X, which is used to standardize the warning index Vw to the range of [0, 1].

5. A road traffic safety dynamic warning system according to claim 4, characterized in that: The warning prompt information is obtained by the following generation method: When the warning index Vw is less than the traffic intersection warning generation threshold YZ, the first-level warning prompt information is obtained, and the relevant prompt slogans of civilized traffic and courtesy to pedestrians are displayed; When the warning index Vw ≥ the traffic intersection warning generation threshold YZ, obtain the secondary warning prompt information, marginally display the relevant prompt slogans of civilized traffic and courtesy to pedestrians, and at the same time remind pedestrians to pay attention to traffic lights and drivers to pay attention to pedestrians; When the warning index Vw ≥ twice the traffic intersection warning generation threshold YZ, the third-level warning prompt information is obtained, and the relevant prompt slogans of civilized traffic and courtesy to pedestrians are displayed on the edge. At the same time, pedestrians are reminded to pay attention to traffic lights and drivers are reminded to pay attention to pedestrians, and the timing of signal lights is adjusted.

6. A road traffic safety dynamic warning system according to claim 5, characterized in that: The signal optimization module includes a signal light timing optimization unit; After the early warning mechanism is triggered, the traffic light timing optimization unit performs adaptive optimization of the timing information according to the traffic light timing at the current traffic intersection, wherein the traffic light timing includes the green light time Tg, the yellow light time Ty and the red light time Tr, and performs fitting calculation with the dynamic adjustment ratio Pa to obtain an optimized traffic light timing data group, including the optimized green light time YTg, the optimized yellow light time YTy and the optimized red light time YTr, and generates a control instruction for the optimized traffic light timing data group to control the traffic light timing at the current traffic intersection; The dynamic adjustment ratio Pa is obtained by the following calculation formula: ； In the formula, γ represents the adjustment coefficient, which is specifically used to adjust the adjustment range of the signal light timing; The optimized green light time YTg is obtained by the calculation formula YTg=Tg*(1+Pa); The optimized yellow light time YTy is obtained by the calculation formula YTy=Ty*(1+Pa); The optimized red light time YTr is obtained by the calculation formula YTr=Tr*(1+Pa).

7. A road traffic safety dynamic warning system according to claim 6, characterized in that: The monitoring and evaluation module includes a flow collection unit and a dynamic adjustment unit; After the early warning mechanism is triggered, the flow collection unit collects pedestrian flow and vehicle flow information at the traffic intersection at a fixed time period, including pedestrian flow Pflow and vehicle flow Vflow; The pedestrian flow Pflow is obtained by the following calculation formula: ； Where Pflow(t) represents the pedestrian flow at time t, Pcount(i, t) represents the count of the i-th pedestrian at time t, and m represents the total number of pedestrians passing through the traffic intersection at time t; The vehicle flow Vflow is obtained by the following calculation formula: ； Where Vflow(t) represents the vehicle flow at time t, Vcount(j, t) represents the count of the jth vehicle at time t, and w represents the total number of vehicles passing the traffic intersection at time t.

8. A road traffic safety dynamic warning system according to claim 7, characterized in that: The dynamic adjustment unit compares and evaluates the pedestrian flow Pflow and the vehicle flow Vflow with the pedestrian flow and the vehicle flow at the traffic intersection before the early warning mechanism is triggered, wherein the pedestrian flow and the vehicle flow at the traffic intersection before the early warning mechanism is triggered are marked as the previous pedestrian flow QPflow and the previous vehicle flow QVflow, and obtains the difference between the pedestrian flow Pflow and the vehicle flow Vflow and the previous pedestrian flow QPflow and the previous vehicle flow QVflow, which are marked as the pedestrian flow change △Pflow and the vehicle flow change △Vflow; Then, the dynamic adjustment coefficient Rad is obtained by calculating the pedestrian flow change △Pflow and the vehicle flow change △Vflow with the pedestrian flow Pflow and the vehicle flow Vflow, and the dynamic adjustment state result of the traffic intersection is obtained, and the early warning mechanism is adaptively adjusted according to the dynamic adjustment state result of the traffic intersection; The dynamic adjustment coefficient Rad is obtained by Obtain calculation formula; The dynamic adjustment state result of the traffic intersection is obtained by the following comparison method: When the dynamic adjustment coefficient Rad>0.1, it means that the traffic flow is increasing, and the signal light timing is adjusted proportionally, including optimizing the green light time YTg, optimizing the yellow light time YTy and optimizing the red light time YTr; When the dynamic adjustment coefficient Rad is less than -0.1, it indicates that the flow rate is decreasing, and the traffic light timing of the initial traffic intersection includes the green light time Tg, the yellow light time Ty and the red light time Tr.

Images