Intelligent traffic risk early warning safe driving guidance method and device and storage medium

Abstract

The invention provides an intelligent traffic risk early warning safe driving guidance method and device and a storage medium, and relates to the technical field of intelligent traffic. The method comprises the following steps: determining a suspected risk event according to multi-source sensing information of a current all-in-one machine; verifying the suspected risk event by adopting adjacent verification, array verification and / or multi-dimensional verification to obtain an effective risk event; according to the effective risk event, generating a dynamic grading and partitioning management and control strategy; wherein the dynamic grading and partitioning management and control strategy is used for applying an induction strategy, a warning strategy and / or an interception strategy for a near zone, a middle zone and a far zone; and when it is detected that the event is removed, generating a partition passage recovery strategy in real time according to the distance between the vehicle and the effective risk event. According to the method, the accuracy of risk event identification is ensured by adopting multiple verification, meanwhile, the dynamic grading and partitioning management and control strategy and the partitioning passing recovery strategy are set, the whole-process closed-loop logic of sensing, monitoring, induction, warning, interception and passing is achieved, and the prevention and control effect is good.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, and in particular to an intelligent transportation risk warning and safe driving guidance method, device and storage medium. Background Technology

[0002] As the highway network continues to expand and become more dense, the proportion of special road sections such as extra-large bridges, long tunnels, sharp bends, steep slopes, and high slopes in the entire network is increasing year by year. This poses multiple challenges to the safe operation of roads. On the one hand, extreme weather (rain, snow, fog, haze, sandstorms) may trigger natural disasters such as rockfalls, mudslides, and road collapses; on the other hand, traffic incidents (rear-end collisions, illegal parking, pedestrian intrusion) can lead to secondary disaster risks.

[0003] In existing technologies, most traffic control systems are primarily reactive and have relatively simple control logic, which makes them prone to false alarms and missed alarms. They are unable to meet the all-weather, all-round safety control needs in complex scenarios, resulting in poor safety control effects and a persistently high rate of secondary accidents. Summary of the Invention

[0004] This invention provides a method, device, and storage medium for intelligent traffic risk warning and safe driving guidance, in order to solve the problem that existing traffic control systems do not meet the safety control needs in complex scenarios and have poor safety control effects.

[0005] In a first aspect, embodiments of the present invention provide a method for intelligent traffic risk warning and safe driving guidance, comprising: Based on the multi-source sensing information from the current all-in-one machine, suspected risk events are identified; Adjacent verification, array verification, and / or multi-dimensional verification are used to verify suspected risk events and obtain valid risk events; Based on valid risk events, a dynamic hierarchical and zoned control strategy is generated; wherein, the dynamic hierarchical and zoned control strategy is used to apply induction strategies, warning strategies and / or interception strategies for near zone, mid zone and far zone; When the event is detected to be resolved, a zoned access restoration strategy is generated in real time based on the distance between the vehicle and the valid risk event.

[0006] Secondly, embodiments of the present invention provide an electronic device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the intelligent traffic risk warning and safe driving guidance method as described in the first aspect or any possible implementation of the first aspect.

[0007] Thirdly, embodiments of the present invention provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the intelligent traffic risk warning and safe driving guidance method as described in the first aspect or any possible implementation thereof.

[0008] This invention provides a method, device, and storage medium for intelligent traffic risk warning and safe driving guidance. The method includes: identifying suspected risk events based on multi-source sensing information from the current integrated device; verifying the suspected risk events using adjacent verification, array verification, and / or multi-dimensional verification to obtain valid risk events; generating a dynamic hierarchical and zonal control strategy based on the valid risk events; wherein the dynamic hierarchical and zonal control strategy is used to apply guidance strategies, warning strategies, and / or interception strategies for near, middle, and far zones; and generating a zonal traffic recovery strategy in real time based on the distance between the vehicle and the valid risk event when the event is detected to be resolved. This application implements a closed-loop logic for the entire process of "perception-monitoring-guidance-warning-interception-reopening," covering the complete stage from risk identification to traffic recovery. First, it uses multi-source sensing information from the integrated device to initially screen suspected risk events. Then, it employs multiple verifications to filter out invalid false alarms. Through multi-source sensing and triple verification, the false alarm rate is significantly reduced, improving the reliability of risk event identification. Simultaneously, it sets up a three-level control strategy of "guidance-warning-interception": the guidance stage provides advance guidance in distant areas to avoid operational risks near risk zones; the warning stage strengthens reminders in the central area to reduce vehicle speed; the interception stage provides precise blocking in the near area for high-level events to prevent the accident from escalating, achieving "differentiated guidance" that balances traffic safety and traffic efficiency; finally, when the event is resolved, a zoned traffic recovery strategy is generated in real time, and control measures are gradually lifted sequentially to ensure a smooth transition of traffic flow to normal. This method forms a complete closed-loop control system, resulting in more precise and effective control, effectively improving road safety and reducing the accident rate. It is suitable for various application scenarios such as major bridges, tunnel entrances and exits, and special areas. Attached Figure Description

[0009] Figure 1 This is a flowchart illustrating the implementation of a smart traffic risk warning and safe driving guidance method provided in an embodiment of the present invention. Figure 2 This is an event verification flowchart provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the intelligent transportation risk warning and safe driving guidance device provided in an embodiment of the present invention. Detailed Implementation

[0010] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0011] See Figure 1 The document illustrates a flowchart of an intelligent traffic risk warning and safe driving guidance method provided by an embodiment of the present invention, which is described in detail below: The aforementioned intelligent transportation risk warning and safe driving guidance methods include: S101: Based on the multi-source sensing information of the current all-in-one machine, identify suspected risk events for event determination.

[0012] The all-in-one device can include radar and surveillance cameras to collect multi-source sensing information and realize real-time monitoring of traffic events (rear-end collisions, illegal parking, driving against traffic, pedestrian intrusion), disaster events (rockfalls, landslides, road collapses, guardrail deformation), and environmental events (heavy rain, fog, dust storms, low light at night), covering multi-dimensional risk scenarios.

[0013] Meanwhile, the all-in-one machine can also collect positioning data, attitude data, environmental data (temperature, humidity, light intensity), meteorological data (wind speed, visibility, precipitation intensity), etc., for subsequent data processing.

[0014] In one possible implementation, the multi-source sensing information includes: radar information and image information; S101 may include: S1011: Target identification event based on radar information; S1012: If a target event is detected, an image of the radar detection area is acquired, and the target event is identified based on the image of the radar detection area; S1013: Match target events identified based on radar information with target events identified based on image recognition; S1014: If a match is found, the target event is determined to be a suspected risk event.

[0015] Millimeter-wave radar scans the road area and, upon detecting targets (vehicles, pedestrians, obstacles) or abnormal conditions (guardrail deformation, road subsidence), records core data such as the target's position, speed, and shape as target events. Simultaneously, image recognition is automatically triggered. A high-definition camera module captures images of the radar detection area, extracting the target's contour features, pixel coordinates, and trajectory. An AI algorithm compares the target events identified based on radar information with those identified based on image recognition. If a match is found (e.g., positional deviation ≤ 5 meters, speed deviation ≤ 0.3 m / s, morphological similarity ≥ 90%), it is identified as a suspected risk event.

[0016] This application adopts a process of "radar first identification trigger, image then secondary verification", which not only uses radar to ensure all-weather coverage of detection, but also uses images to make up for the ambiguity of radar target identification, which greatly reduces the false judgment rate and false negative rate of a single sensor and effectively improves the detection accuracy.

[0017] S102: Use adjacent verification, array verification and / or multi-dimensional verification to verify suspected risk events and obtain valid risk events; To improve the accuracy of detection and ensure the accuracy of risk event location, this application also sets up a multi-level verification mechanism.

[0018] In one possible implementation, refer to Figure 2 S102 may include: S1021: Verify suspected risk events using adjacent all-in-one machines that are in the same direction and on the same side as the current all-in-one machine; S1022: If the verification passes, the suspected risk event is determined to be a valid risk event; The system retrieves synchronous sensing data (event timestamps, location coordinates, target type, etc.) from adjacent devices on the same side and in the same direction as the current all-in-one device. It then checks whether the events from these adjacent devices match the suspected risk events of the current all-in-one device. For example, if similar events are detected within the same time window (±2 seconds) and at similar locations (±5 meters), the verification passes if at least one adjacent all-in-one device detects a matching event, confirming the suspected risk event as a valid risk event. Otherwise, it proceeds to the next level of verification.

[0019] S1023: If the verification fails, a device array formed by adjacent all-in-one machines on opposite sides in the same direction will be used to verify the suspected risk event; S1024: If the verification passes, the suspected risk event is determined to be a valid risk event; If the adjacent all-in-one machine on the same side does not detect the event or the data matching is insufficient, the second level of verification is initiated. Adjacent all-in-one machines on opposite sides in the same direction are selected to form a horizontal and vertical device array, creating three-dimensional sensing coverage, and multi-source data consistency verification is performed. If the array data calculation result is consistent with the event information of the previous level of verification (positional deviation ≤ 0.5 meters), the verification is considered successful; otherwise, the third level of verification is initiated.

[0020] Simultaneously, based on the principle of triangulation, the event coordinate data of all devices in the array can be fused together, and combined with the road lane width and marking position, to achieve lane-level positioning calculation of the precise location of events (lane level, error ≤ 0.5 meters). Furthermore, through multi-device data fusion, the scope of event impact can be verified (such as whether falling rocks cover multiple lanes), providing a precise basis for the subsequent delineation of the scope of zoned control, and avoiding over- or under-control due to positioning deviation of a single device.

[0021] S1025: If the verification fails, multi-dimensional regional perception data will be integrated to verify the suspected risk event; S1026: If the verification passes, the suspected risk event is determined to be a valid risk event; S1027: Otherwise, determine that the suspected risk event is not a valid risk event.

[0022] Finally, a wider range of "multi-dimensional regional perception data" is integrated, including but not limited to: pre-level verification data, environmental data (temperature, humidity, road surface condition), meteorological data (wind speed, visibility, precipitation type), and historical data (occurrence rate of similar events in the region, handling records); the multi-dimensional data is fused and analyzed through algorithms to conduct final verification of suspected risk events.

[0023] In one possible implementation, S1025 may include: 1. Acquire multi-dimensional sensing data, environmental data, meteorological data, and historical data; 2. Determine the perception verification score, environmental verification score, meteorological verification score, and historical verification score based on multi-dimensional perception data, environmental data, meteorological data, and historical data respectively; 3. The comprehensive verification score is obtained by weighted summation of the perception verification score, environmental verification score, meteorological verification score, and historical verification score; 4. If the overall verification score is greater than the preset score, the verification is considered passed; otherwise, the verification fails.

[0024] The system acquires multi-dimensional perception information from the pre-detection stage (e.g., target features and motion status from radar and video), environmental data (e.g., temperature, humidity, road surface condition (whether there is water accumulation, etc.), meteorological data (wind speed, visibility, precipitation type (rain / snow / fog)), and historical data (frequency of similar events in the area over the past year, and handling records). It calculates perception verification scores, environmental verification scores, meteorological verification scores, and historical verification scores, which represent the degree of support of each data point for the probability of an event occurring. The system weights and sums the data based on the importance of each data dimension to the event verification to obtain the final comprehensive verification score, determining whether the verification passes. This effectively improves the accuracy of complex scenarios (such as determining slow-moving vehicles in heavy rain).

[0025] For example, the weights corresponding to the perception verification score, environmental verification score, meteorological verification score, and historical verification score can be 30%, 25%, 25%, and 20%, respectively.

[0026] For example, the preset score can be 80 points. If the comprehensive verification score is greater than 80 points, the verification passes. Furthermore, if the comprehensive verification score is less than 60 points, it is determined to be a suspected false alarm and requires further manual review. If the comprehensive verification score is between 60 and 79 points, it is determined to be an "event to be observed," and re-verification is performed after 10 seconds of continuous monitoring.

[0027] This application balances efficiency and accuracy by moving from "close-range rapid verification" to "wide-range multi-dimensional verification," avoiding waste of resources.

[0028] S103: Generate dynamic hierarchical and zonal control strategies based on valid risk events; wherein, dynamic hierarchical and zonal control strategies are used to apply induction strategies, warning strategies and / or interception strategies for near zone, mid zone and far zone. When vehicles are traveling on the road, there is an incompressible reaction distance, braking distance, and space required for lane changing / diversion. Therefore, this application automatically generates a three-tiered control strategy of "far-area guidance, mid-area warning, and near-area interception" based on effective risk events, so as to achieve precise, layered, and efficient handling of risk events and minimize the impact of events on traffic order and safety.

[0029] In one possible implementation, S103 may include: S1031: Obtain the severity, scope of impact, location, weather conditions, and road topology of valid risk events; S1032: Determine the event level based on its severity and scope of impact; Based on the scope of the incident's impact (whether it blocks lanes, whether it threatens structural safety) and its severity (whether immediate interception is required), the incident level is determined as follows: Emergency (landslides, multi-vehicle rear-end collisions, etc.), Important (guardrail deformation, rockfalls, etc.), Caution (driving slowly, minor scrapes, severe weather, etc.). S1033: Determine the event control zones based on location and road topology; This application, based on the location of the incident and the road topology, delineates the incident control zones into a near zone (the area close to the incident center), a middle zone (the area at a medium distance from the incident center), and a far zone (the area far from the incident center).

[0030] Regarding the event control partitioning rules, in one possible implementation, the event control partitions include: near zone, middle zone, and far zone; S1033 may include: 1. Determine the distance between zones based on the weather conditions; 2. Determine the distance between the current vehicle and the valid risk event based on its location; 3. Determine the event control zone based on the distance between the current vehicle and the valid risk event, and the zone distance.

[0031] Different weather conditions (such as sunny, rainy, foggy, and snowy weather) directly affect vehicle braking distance, driver visibility, and road traffic efficiency. Therefore, different distance thresholds for near, middle, and far zones are preset for different weather scenarios. The actual spatial distance between the current vehicle and the effective risk event is used as the basis for judgment. Combined with the uniqueness of the passage path implied by the road topology, the accuracy and effectiveness of distance calculation are ensured, and the accurate determination of the control zone to which the current vehicle belongs is ultimately achieved.

[0032] The three-tiered zoning of near, middle, and far zones forms a gradient control architecture from the core response area to the outer early warning area. This architecture supports the implementation of different control measures such as emergency braking, deceleration reminders, and path planning at different levels, thereby achieving layered risk mitigation and improving the effectiveness and scientific nature of overall control.

[0033] S1034: Generate dynamic hierarchical and zonal control strategies based on weather scenarios, event levels, and event control zones.

[0034] Different weather conditions correspond to different road capacity and vehicle handling limits; event level reflects the severity and scope of the risk source itself, serving as the basic basis for determining strategy strength; event control zoning clarifies the spatial relationship between the current vehicle and the risk source, directly corresponding to the strategy execution priority and specific measures. This application deeply couples the above three dimensions, pre-setting a corresponding set of control measures for each combination, generating a dynamic hierarchical zoning control strategy that is targeted, adaptable, and tiered.

[0035] In one possible implementation, the dynamic hierarchical zoning control strategy may include: a far-area guidance strategy, a mid-area warning strategy, and a near-area interception strategy; S1034 may include: 1. When the event control zone is a remote zone, generate a remote zone guidance strategy based on the weather scenario and event level; 2. When the event control zone is the middle zone, generate a middle zone warning strategy based on the weather scenario and event level; 3. When the event control zone is the near zone, generate a near zone interception strategy based on the weather scenario and event level.

[0036] Advance guidance for remote areas: When vehicles enter remote areas, the terminal equipment in remote areas respond first: gantry information boards and roadside guidance screens release scenario-based and hierarchical guidance information, and the navigation platform pushes road conditions and detour suggestions; at the same time, it guides vehicles to adjust their driving status in advance (slow down, change lanes, queue) to avoid secondary accidents caused by sudden braking or sudden lane changes when approaching risk areas.

[0037] Central Zone Enhanced Reminders: After a vehicle enters the central zone, the intelligent perception and guidance integrated machine will provide enhanced reminders through light warnings (yellow flashing / red and blue alternating flashing) and voice broadcasts (risk type + avoidance suggestions); the roadside screen displays distance prompts and operation instructions to further reduce vehicle speed and improve driver reaction time.

[0038] Precise blocking in the vicinity: After a vehicle enters the vicinity, for emergency and important incidents, physical interception (soft barrier device) and strong warning (red no-entry light + high-frequency voice) are activated; the traffic flow is forcibly blocked from entering the core risk area to prevent the accident from escalating and to buy time for incident handling.

[0039] This application implements corresponding control measures according to the event control zones and event levels. The emergency level focuses on "near-area interception + mid-area warning + far-area diversion", the important level focuses on "near-area avoidance guidance + mid-area deceleration reminder + far-area warning", and the attention level focuses on "smooth guidance throughout the entire area".

[0040] For example, the control logic based on event level and event control partition is shown in Table 1.

[0041] Table 1. Comparison of Control Logic between Event Levels and Event Control Zones

[0042] Since the weather scenario is dynamically adjusted, the dynamic hierarchical and zonal control strategy is adjusted based on Table 1.

[0043] For example, on a sunny day: with "clear guidance" as the core, regular light warnings (such as red for no entry, yellow for slow down) are activated, along with concise voice prompts, and the information board displays complete event information (such as "rockfall 500 meters ahead, slow down"). Rain / snow weather: Enhance "anti-slip + warning", add "road surface is slippery, keep a safe distance" prompt to the voice broadcast, and increase the font size of the guidance screen and use high brightness display to avoid rain reflection affecting recognition; In foggy / hazy / dusty weather: The core of the system is "outline indication + close-range interception". Activate the light barrier indication soft barrier (forming a horizontal light strip to clearly define the lane boundary), increase the frequency of voice prompts to once every 3 seconds, and reduce the distance of the far-area guidance (e.g., from 1000 meters to 500 meters) to prevent drivers from missing the warning due to low visibility. If no incident occurs, activate the yellow light flashing to remind drivers to proceed slowly and safely. At night: Increase the brightness of warning lights (e.g., increase the brightness of red no-entry lights by 50%), turn off unnecessary bright lights (e.g., avoid direct light affecting visibility), and use low-frequency tones for voice prompts to reduce interference with nearby residents.

[0044] Once the incident has been handled (such as clearing the landslide and towing away the accident vehicles), the traffic restoration process will be initiated after manual confirmation or automatic system detection.

[0045] S104: When the event is detected to be resolved, generate a zoned access restoration strategy in real time based on the distance between the vehicle and the valid risk event.

[0046] In one possible implementation, S104 may include: S1041: Determine the event recovery zone in real time based on the distance between the vehicle and the valid risk event, as well as the zone distance; S1042: Generate a zone access recovery strategy based on the event recovery zone and weather scenario.

[0047] After the incident is handled, the system gradually restores traffic flow according to the "scenario adaptation, near-medium-far sequence" to avoid secondary risks caused by sudden changes in traffic flow. For example, in the near area: first, disable the no-entry and interception equipment (such as red lights and soft barriers), switch to green traffic lights + voice "Risk cleared, traffic resumed", and simultaneously detect the on-site traffic flow status; Central Zone: After the traffic flow in the nearby area begins to move smoothly, turn off the warning devices (such as red and blue lights, yellow flashing lights) and switch to green traffic lights + voice prompts. The voice prompt will say "Risk cleared, drive smoothly" to guide the queued vehicles to start smoothly. Remote areas: Once traffic flow in the central area returns to normal speed, the information board will be updated to green passage information (such as "Risk ahead cleared, normal passage"), the navigation platform will cancel the detour prompts, and guide vehicles back to their original routes; Scene adaptation adjustments: For example, when traffic resumes at night, maintain the brightness of the green traffic light until the traffic flow stabilizes; when traffic resumes in foggy weather, extend the on-time of the outline lights to ensure that drivers adapt to changes in road conditions.

[0048] Once the control strategy is generated, the information dissemination and guidance mechanism of the dynamic hierarchical and zoning control strategy can adopt three different modes.

[0049] 1. Automatic guidance + timed update mode: Suitable for scenarios where the event status is stable and does not require frequent adjustments (such as temporary construction or continuous fog). The system automatically generates dynamic hierarchical and zoned control strategies and updates the content (such as "Construction 2 hours remaining") at preset time intervals (such as 5-10 minutes, which can be customized) until the event ends. The information is synchronously pushed to the gantry information board, roadside guidance screen and navigation platform to ensure consistency of information across multiple terminals.

[0050] 2. Automatic guidance + manual confirmation mode: Suitable for scenarios with complex events requiring manual judgment (such as multi-vehicle accident scenes or suspected landslides). The system automatically generates dynamic hierarchical and zoned control strategies and pushes them to management personnel terminals (such as monitoring center screens or mobile apps). After confirming the rationality of the strategy based on on-site video and feedback from maintenance personnel, the management personnel click to publish. The information is updated every 10 minutes. If the event status changes (such as the landslide area expands), it can be adjusted in real time and re-confirmed for publication, balancing the efficiency of automation with the accuracy of human decision-making.

[0051] 3. Automatic guidance + emergency trigger mode: Applicable to emergency events (such as sudden landslides, multi-vehicle rear-end collisions). After the system detects the event and passes multiple verifications, it immediately issues dynamic hierarchical and zoned control policies (such as activating red traffic lights, soft barrier devices, and voice broadcasts) without manual confirmation, and pushes alarm information to management personnel at the same time. After the information is issued, the system automatically checks the event status every 3 minutes. If the risk increases (such as an increase in falling rocks), it automatically upgrades control measures (such as expanding the interception range) to ensure rapid response to high-risk scenarios.

[0052] Information presentation adopts a unified standard of "color-icon-content", and enhances recognizability by combining scenarios and hierarchical classification: Emergency level: Red text + "No Entry" icon, concise and clear content (e.g., "No Entry Ahead, Detour Immediately"), flashing display in foggy / nighttime conditions; Importance level: Red text + "Slow down" icon, the content includes the risk type and distance (e.g., "Rockfall 300 meters ahead, slow down"). Caution Level: Yellow text + "Caution" icon, with content focusing on guidance and suggestions (such as "Construction ahead, merge into the right lane"). Level 1: Green text + “Pass” icon, with clear indication of recovery (e.g., “Risk cleared, normal passage”).

[0053] This application establishes a closed-loop logic encompassing the entire process of "perception-monitoring-control-traffic," covering the complete stage from risk identification to traffic recovery. The perception phase collects multi-source basic data; the monitoring phase ensures the authenticity of events through multiple verifications; the control phase includes guidance, warnings, and interception; the guidance phase provides advance guidance to distant areas, avoiding operational risks when near risk zones; the warning phase strengthens alerts in the central area and reduces vehicle speed; the interception phase achieves precise blocking in the near area for high-level events, preventing the accident from escalating; and the traffic recovery phase gradually lifts controls in the order of "near-medium-far," ensuring a smooth transition of traffic flow to normal. The entire process operates automatically without human intervention, while also supporting manual confirmation of risk clearance, forming an "automated + controllable" closed-loop control system.

[0054] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0055] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0056] Figure 3 A schematic diagram of the intelligent traffic risk warning and safe driving guidance device provided in an embodiment of the present invention is shown. For ease of explanation, only the parts related to the embodiment of the present invention are shown, and are described in detail below: like Figure 3 As shown, the intelligent transportation risk warning and safe driving guidance device includes: Event recognition module 21 is used to identify suspected risk events based on the multi-source sensing information of the current all-in-one machine; Verification module 22 is used to verify suspected risk events by using adjacent verification, array verification and / or multi-dimensional verification to obtain valid risk events; The event management module 23 is used to generate dynamic hierarchical and zonal management strategies based on valid risk events; wherein, the dynamic hierarchical and zonal management strategies are used to apply induction strategies, warning strategies and / or interception strategies for near zone, mid zone and far zone. The access restoration module 24 is used to generate a zoned access restoration strategy in real time based on the distance between the vehicle and the valid risk event when the event is detected to be resolved.

[0057] In one possible implementation, the multi-source sensing information includes: radar information and image information; the event recognition module 21 may include: The first identification unit is used to identify target events based on radar information; The second identification unit is used to acquire an image of the radar detection area and identify the target event based on the image of the radar detection area if a target event is detected. The matching unit is used to match target events obtained based on radar information recognition with target events obtained based on image recognition; The event determination unit is used to determine the target event as a suspected risk event if a match is found.

[0058] In one possible implementation, the verification module 22 may include: The first verification unit is used to verify suspected risk events using adjacent all-in-one machines on the same side and in the same direction as the current all-in-one machine. The first judgment unit is used to determine the suspected risk event as a valid risk event if the verification passes. The second verification unit is used to verify the suspected risk event by using an array of devices formed by adjacent all-in-one machines on the opposite side of the same direction if the verification fails. The second judgment unit is used to determine the suspected risk event as a valid risk event if the verification passes. The third verification unit is used to verify suspected risk events by integrating multi-dimensional regional perception data if the verification fails. The third judgment unit is used to determine the suspected risk event as a valid risk event if the verification passes. The fourth judgment unit is used to determine otherwise that the suspected risk event is not a valid risk event.

[0059] In one possible implementation, the third verification unit can be specifically used for: 1. Acquire multi-dimensional sensing data, environmental data, meteorological data, and historical data; 2. Determine the perception verification score, environmental verification score, meteorological verification score, and historical verification score based on multi-dimensional perception data, environmental data, meteorological data, and historical data respectively; 3. The comprehensive verification score is obtained by weighted summation of the perception verification score, environmental verification score, meteorological verification score, and historical verification score; 4. If the overall verification score is greater than the preset score, the verification is considered passed; otherwise, the verification fails.

[0060] In one possible implementation, the event management module 23 includes: The event parameter acquisition unit is used to acquire the severity, scope of impact, location, weather scenario, and road topology of valid risk events; The rating output unit is used to determine the event rating based on its severity and scope of impact. The first partition output unit is used to determine the event control partition based on location and road topology; The first strategy output unit is used to generate dynamic hierarchical and zonal control strategies based on weather scenarios, event levels, and event control zones.

[0061] In one possible implementation, the event control partition includes: a near zone, a middle zone, and a far zone; the first partition output unit can be specifically used for: 1. Determine the distance between zones based on the weather conditions; 2. Determine the distance between the current vehicle and the valid risk event based on its location; 3. Determine the event control zone based on the distance between the current vehicle and the valid risk event, and the zone distance.

[0062] In one possible implementation, the first strategy output unit can be specifically used for: 1. When the event control zone is a remote zone, generate a remote zone guidance strategy based on the weather scenario and event level; 2. When the event control zone is the middle zone, generate a middle zone warning strategy based on the weather scenario and event level; 3. When the event control zone is the near zone, generate a near zone interception strategy based on the weather scenario and event level.

[0063] In one possible implementation, the access restoration module 24 may include: The second partition output unit is used to determine the event recovery partition in real time based on the distance between the vehicle and the valid risk event, as well as the partition distance; The second strategy output unit is used to generate a zone access recovery strategy based on the event recovery zone and weather scenario.

[0064] This invention also provides an electronic device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described in the above method embodiments. Exemplarily, the electronic device can be a host computer, but this is not limited thereto.

[0065] This invention also provides an intelligent traffic risk early warning system.

[0066] The following will provide a detailed explanation using specific scenarios.

[0067] (I) Extra-large bridges The bridge has a large span and complex structure, making it susceptible to strong winds, floods, and ship collisions, posing risks of bridge deck collapse and guardrail deformation. Passenger vehicles, dangerous goods vehicles, and heavy vehicles pass through frequently, making traffic management difficult after an accident and increasing the risk of secondary accidents. The bridge has no obstructions on both sides, resulting in low visibility in rainy or foggy weather and limiting the driver's vision.

[0068] System deployment plan: Integrated cameras are installed on both sides of the bridge railing at intervals of 20-30 meters (for both lanes), with the intervals increased to 15 meters in key areas (such as near the piers and at the joints of the bridge deck); a mini weather station is installed every 500 meters to monitor wind speed, precipitation, and visibility; and pan-tilt remote-controlled cameras are added at both ends of the bridge to enhance video verification capabilities.

[0069] One intelligent communication base station is deployed every 800-1000 meters, using CAN bus wiring along the guardrail, with LoRa wireless as backup, to ensure communication coverage throughout the entire bridge without any dead zones.

[0070] Remote node controllers are deployed near the midpoint of the bridge, and the system software platform uses a special rule library for "bridge structural risk". A "speed limit + no overtaking" control strategy is customized for strong wind scenarios, and "slow down + keep distance" voice prompts are enhanced for bridge water accumulation scenarios.

[0071] A gantry information board is deployed 1,000 meters before the bridge entrance, and roadside guidance screens are installed every 300 meters on the bridge deck. Soft barriers are installed near the bridge piers. The light grid outline function is activated in foggy weather, and the red and blue lights flash frequently with the voice "Strong wind warning, drive carefully" is activated in strong wind weather.

[0072] Functions: Real-time monitoring of bridge structural anomalies (such as guardrail deformation and bridge deck settlement) and traffic incidents (such as rear-end collisions), ensuring accurate risk identification through multiple verifications; issuing scenario-based guidance information for severe weather and key vehicles; and quickly initiating nearby interception after an accident to prevent secondary accidents and ensure bridge traffic safety.

[0073] (ii) Tunnel entrances and exits The significant difference in light levels inside and outside tunnels can cause drivers to experience blind spots due to the need for light adaptation, increasing the risk of collisions. The entrances and exits are located near mountains, making them susceptible to landslides and rockfalls during heavy rain or earthquakes. The enclosed space of tunnels makes it difficult to evacuate traffic after an accident, increasing the risk of congestion and secondary accidents. At night or in foggy weather, visibility at the entrances and exits is poor, making it easy for vehicles to drift out of their lanes.

[0074] System deployment plan: Integrated monitoring units are deployed at 30-meter intervals within a 1000-2000-meter range before and after the tunnel entrance and exit, with the intervals near the mountain slopes increased to 20 meters. Inside the tunnel, one integrated monitoring unit is deployed every 30 meters to achieve seamless connection between internal and external monitoring. Miniature weather stations are deployed within 50 meters of the entrance and exit to focus on monitoring fog and rainstorms.

[0075] A smart communication base station is deployed 100-200 meters from the tunnel entrance and exit. CAN bus wired communication is used inside the tunnel (to prevent interference), and LoRa wireless backup is enabled outside the tunnel. The base station is linked with the tunnel monitoring system to share video data for radar verification.

[0076] Remote node controllers and system software platforms are deployed at the tunnel management station, and a strategy library for "tunnel entrance and exit scenarios" is enabled. To address the pain point of "light and darkness adaptation", a transition strategy of "gradually reducing the brightness of the guidance screen before the entrance and gradually increasing the brightness before the exit" is customized. To address the risk of falling rocks, a "slope displacement threshold warning" rule is set.

[0077] Information boards are installed on the tunnel entrance and exit gantries, high-volume voice broadcasting equipment is deployed on the roadside, and light barriers and soft barriers are installed in the nearby area (within 50 meters of the entrance and exit); the guidance lights are activated in a constant-on mode at night, and a triple warning of "light barriers + voice + information boards" is activated in foggy weather; rockfall incidents trigger a red no-entry zone + soft barrier in the nearby area.

[0078] Functions: Real-time monitoring of natural disasters such as landslides and rockfalls, as well as traffic incidents inside and outside the tunnel; mitigation of the "light-dark adaptation" problem through scenario-based strategies; rapid activation of graded and zoned control after an accident to guide traffic flow evacuation, avoid congestion and secondary accidents, and ensure safe passage at tunnel entrances and exits.

[0079] (III) Special areas (sharp bends, high slopes, areas prone to fog) Sharp bends obstruct visibility, making it easy for vehicles to occupy lanes or speed, leading to collisions; high slopes have unstable geology, making them prone to landslides and rockfalls; visibility drops sharply in areas with frequent fog (as low as less than 5 meters), easily causing multi-vehicle pile-up collisions; traditional equipment has low perception accuracy in low visibility and strong interference environments.

[0080] System deployment plan: At sharp bends, at the bottom of high slopes, and in areas with frequent fog, integrated weather stations are densely deployed at intervals of 20-30 meters; attitude sensors are added to the top of high slopes to monitor slope displacement; and a miniature weather station is deployed every 500 meters in fog areas to monitor visibility in real time.

[0081] One smart communication base station is deployed every 800-1000 meters, using LoRa wireless communication as the primary method (to adapt to the difficulty of wiring in complex terrain) and CAN bus as a secondary method (for key areas); the base station signal coverage ensures stable data interaction between adjacent integrated units and meets the requirements for adjacent verification.

[0082] Remote node controllers are deployed at nearby maintenance stations, and the system software platform enables the "Special Area Special" strategy library. For sharp bend scenarios, a strategy of "activating yellow strobe light + voice 'Sharp bend ahead, slow down'" is customized. For fog scenarios, the rule of "automatically triggering light barrier soft blocking when visibility is <50 meters" is enabled.

[0083] Lane-level guidance screens are installed at sharp bends, voice broadcasting equipment is installed at the bottom of high slopes, and light barriers are deployed near fog areas; the brightness of warning lights on bends is enhanced at night, and red and blue lights flash alternately with the voice message "Landslide ahead, no passage" is triggered when a landslide risk is detected.

[0084] Function: Through high-density sensing and quadruple verification, it accurately identifies geological disasters and traffic incidents in complex environments; it customizes control strategies based on scene characteristics, guides vehicles to drive in a standardized manner on sharp bends, strengthens the outline and interception in foggy areas, and provides early warning of landslide risks on high slopes, thus comprehensively improving driving safety in special areas.

[0085] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Unless otherwise specified or in conflict with logic, the terminology and / or descriptions between different embodiments are consistent and can be referenced interchangeably. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0086] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A smart traffic risk pre-warning safe driving and route guiding method, characterized in that, The method comprises the following steps: According to the multi-source perception information of the current integrated machine, a suspected risk event is determined; The suspected risk event is verified by using adjacent verification, array verification and / or multi-dimensional verification to obtain an effective risk event; According to the effective risk event, a dynamic hierarchical partitioned control strategy is generated; wherein the dynamic hierarchical partitioned control strategy is used to apply an induction strategy, a warning strategy and / or an interception strategy for the near zone, the middle zone and the far zone; When an event is detected to be resolved, a partitioned traffic recovery strategy is generated in real time according to the distance between the vehicle and the effective risk event. 2.The intelligent traffic risk pre-warning and safe driving guidance method according to claim 1, characterized in that, The multi-source perception information includes radar information and image information; the suspected risk event is determined according to the multi-source perception information of the current integrated machine, which comprises the following steps: A target event is identified according to the radar information; If the target event is identified, the image of the radar detection area is obtained, and the target event is identified according to the image of the radar detection area; The target event identified based on the radar information and the target event identified based on the image are matched; If the match is consistent, the target event is determined to be the suspected risk event. 3.The intelligent traffic risk pre-warning and safe driving guidance method according to claim 2, characterized in that, The suspected risk event is verified by using adjacent verification, array verification and / or multi-dimensional verification to obtain an effective risk event, which comprises the following steps: The suspected risk event is verified by using the same direction and same side adjacent integrated machine of the current integrated machine; If the verification is passed, the suspected risk event is determined to be the effective risk event; If the verification is not passed, the suspected risk event is verified by using the device array formed by the same direction and opposite side adjacent integrated machine; If the verification is passed, the suspected risk event is determined to be the effective risk event; If the verification is not passed, the suspected risk event is verified by fusing multi-dimensional regional perception data; If the verification is passed, the suspected risk event is determined to be the effective risk event; Otherwise, it is determined that the suspected risk event is not the effective risk event. 4.The intelligent traffic risk pre-warning and safe driving guidance method according to claim 3, characterized in that, The suspected risk event is verified by fusing multi-dimensional regional perception data, which comprises the following steps: Multi-dimensional perception data, environmental data, meteorological data and historical data are obtained; The perception verification score, the environmental verification score, the meteorological verification score and the historical verification score are determined according to the multi-dimensional perception data, the environmental data, the meteorological data and the historical data respectively; The perception verification score, the environmental verification score, the meteorological verification score and the historical verification score are weighted and summed to obtain a comprehensive verification score; If the comprehensive verification score is greater than a preset score, the verification is passed; otherwise, the verification is not passed. 5.The intelligent traffic risk pre-warning and safe driving guidance method according to any one of claims 1 to 4, characterized in that, The dynamic hierarchical partitioned control strategy is generated according to the effective risk event, which comprises the following steps: The severity, influence range, location, weather scenario and road topology of the effective risk event are obtained; The event level is determined according to the severity and the influence range; The event control partition is determined according to the location and the road topology; The dynamic hierarchical partitioned control strategy is generated according to the weather scenario, the event level and the event control partition. 6.The intelligent traffic risk pre-warning and safe driving guidance method according to claim 5, characterized in that, The event control zones include: near zone, middle zone, and far zone; determining the event control zones based on the location and the road topology includes: Determine the zone distance based on the described weather scenario; The distance between the current vehicle and the effective risk event is determined based on the location; The event control zone is determined based on the distance between the current vehicle and the valid risk event, and the zone distance. 7.The intelligent traffic risk pre-warning and safe driving guidance method according to claim 6, characterized in that, The step of generating the dynamic hierarchical partition control strategy based on the weather scenario, the event level, and the event control partition includes: When the event control zone is a remote zone, the remote zone guidance strategy is generated based on the weather scenario and the event level. When the event control zone is the middle zone, the middle zone warning strategy is generated based on the weather scenario and the event level; When the event control zone is the near zone, the near zone interception strategy is generated based on the weather scenario and the event level. 8.The intelligent traffic risk pre-warning and safe driving guidance method according to claim 6, characterized in that, The real-time generation of zoned traffic recovery strategies based on the distance between the vehicle and the effective risk event includes: The event recovery zone is determined in real time based on the distance between the vehicle and the valid risk event, and the zone distance; Based on the event recovery partition and the weather scenario, generate the partition access recovery strategy.

9. An electronic device, comprising: It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the intelligent traffic risk warning and safe driving guidance method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the intelligent traffic risk warning and safe driving guidance method as described in any one of claims 1 to 8.

Images