Abnormal accident early warning method, system and device and electronic equipment

Abstract

The embodiment of the invention relates to the technical field of vehicles, and discloses an abnormal accident early warning method, system and device and electronic equipment, and the method comprises the steps: responding to the received abnormal accident information sent by a vehicle, and carrying out the early warning of an abnormal accident based on the position of the abnormal accident indicated in the abnormal accident information and the accident type of the abnormal accident; determining a target area influenced by the abnormal accident; the abnormal accident information comprises an accident type of an abnormal accident, a position of the abnormal accident and image data of the abnormal accident; the accident type comprises at least one of road collapse, bridge floor fracture, flood, landslide and rockfall; generating early warning information based on the type of the abnormal accident, the position of the abnormal accident and the image data of the abnormal accident; and sending the early warning information to an adjacent vehicle, wherein the distance between the adjacent vehicle and the target area is smaller than or equal to a preset distance range. By applying the technical scheme of the invention, early warning can be quickly carried out when an abnormal accident occurs.

Description

Technical Field

[0001] This application relates to the field of vehicle technology, specifically to an abnormal accident early warning method, system, device, and electronic equipment. Background Technology

[0002] With the continuous improvement of highway construction, more and more sections of roads traverse complex geological areas such as mountains and rivers. This increases the complexity and risk of the road traffic environment, directly leading to a frequent occurrence of sudden and abnormal accidents caused by natural disasters, such as road collapses, bridge collapses, landslides, floods, and rockfalls. These accidents are characterized by their sudden occurrence and wide-ranging impact, easily leading to chain-reaction traffic accidents and posing a serious threat to the safety of passing vehicles. Furthermore, existing early warning methods largely rely on manual patrols or post-event feedback, resulting in delayed response times and limited coverage, failing to issue timely warnings. Therefore, how to quickly issue early warnings in the event of abnormal accidents is a pressing issue that needs to be addressed. Summary of the Invention

[0003] In view of the shortcomings of the prior art, the purpose of this application is to provide an abnormal accident early warning method, system, device and electronic device, which aims to achieve rapid early warning when an abnormal accident occurs.

[0004] In a first aspect, embodiments of this application provide an abnormal accident early warning method, the method comprising: responding to receiving abnormal accident information sent by a vehicle, determining a target area affected by the abnormal accident based on the location and accident type of the abnormal accident indicated in the abnormal accident information; the abnormal accident information includes: the accident type of the abnormal accident, the location of the abnormal accident, and image data of the abnormal accident; the accident type includes at least one of the following: road collapse, bridge deck fracture, flood, landslide, rockfall; generating early warning information based on the type of the abnormal accident, the location of the abnormal accident, and the image data of the abnormal accident; the early warning information is used to prompt other vehicles passing through the target area to take evasive action; and sending the early warning information to nearby vehicles, wherein nearby vehicles are vehicles whose distance from the target area is less than or equal to a preset distance range.

[0005] The beneficial effects of this application are as follows: By receiving abnormal accident information reported by vehicles and combining the location, type, and image data of the abnormal accidents, the target area is accurately determined, ensuring that the early warning focuses only on the actually affected area. Based on key accident information, the early warning information is sent to nearby vehicles, achieving accurate delivery of the warning information to the target vehicles. This avoids interference caused by irrelevant vehicles receiving the warning information and ensures that vehicles within the affected area receive the warning content in a timely manner. Compared to existing technologies that obtain abnormal accident information through manual labor and fixed roadside equipment, this method, by having vehicles directly report information as sensing devices, eliminates the need for manual patrols or periodic acquisition of abnormal information by roadside equipment. This avoids the time and space limitations of fixed sensing methods, improves the timeliness and coverage of accident information acquisition, and thus shortens the early warning time for abnormal accidents.

[0006] In some embodiments, the abnormal accident information further includes: vehicle attitude information; determining the target area affected by the abnormal accident based on the location and type of the abnormal accident indicated in the abnormal accident information, including: determining the area within a preset distance from the location of the abnormal accident as the initial area of ​​the abnormal accident; inputting the vehicle attitude information into an abnormal accident severity determination model, and determining the degree of deviation of the vehicle from its normal driving posture based on the vehicle attitude information to determine the severity of the abnormal accident; determining the degree of impact of the abnormal accident based on the type of the abnormal accident; determining the area adjustment change amount based on the severity and degree of impact of the abnormal accident; the area adjustment change amount is positively correlated with the severity; the area adjustment change amount is positively correlated with the degree of impact; adjusting the initial area based on the area adjustment change amount to obtain the target area.

[0007] Based on the above technical means, the initial area is determined based on the accident location, the severity of the accident is determined by the vehicle posture information, the degree of impact is determined by the accident type, and the area range is adjusted positively according to the severity and degree of impact to finally obtain a target area that fits the actual risk, avoiding the failure of the early warning due to the area being too large or too small.

[0008] In some embodiments, the target area is determined by: extracting spatial distribution features of the abnormal accidents from the image data of the abnormal accidents in the abnormal accident information; the spatial distribution features include at least one of the following: the scope of the accident's impact and the degree of road surface damage; determining the basic impact area of ​​the abnormal accident based on the scope of the accident's impact and the accident type of the abnormal accident; determining the degree of risk of the abnormal accident causing potential accidents based on the degree of road surface damage; and correcting the basic impact area based on the degree of risk of causing potential accidents to obtain the target area.

[0009] Based on the aforementioned technical means, two major spatial features—the scope of the accident's impact and the degree of road surface damage—are extracted from abnormal accident images. First, a basic impact area is determined based on the scope of impact and the type of accident. Then, potential accident risks are predicted based on the degree of road surface damage. Finally, the basic area is corrected according to the risk level. This approach not only considers the direct impact of the current accident but also predicts potential risks based on the degree of road surface damage, thereby improving the overall accuracy of the target area.

[0010] In some embodiments, sending warning information to nearby vehicles includes: determining a target geographic range subscription channel corresponding to the abnormal accident based on the location of the abnormal accident; the geographic range subscription channel is a message transmission channel for sending warning information; each geographic range subscription channel corresponds to a geographic range of a fixed road segment; the warning information is published to the target geographic range subscription channel; the target geographic range subscription channel is used to send warning information to nearby vehicles that have subscribed to the target geographic range subscription channel.

[0011] Based on the aforementioned technical methods, vehicles dynamically subscribe to geographical channels corresponding to the current road segment during operation, eliminating the need for continuous location reporting to the cloud for alerts. This significantly reduces the communication load between vehicles and the cloud. Furthermore, the Broker node only needs to publish the alert information to the corresponding channel to automatically reach all subscribed vehicles, avoiding the complex calculations of location filtering for all vehicles in the network and improving alert distribution efficiency. Simultaneously, the strong binding between channels and fixed road ranges fundamentally prevents invalid pushes of alert information to vehicles in unaffected areas, reducing information interference to onboard terminals and lowering the transmission pressure on cloud servers.

[0012] In some embodiments, sending warning information to nearby vehicles includes: determining the duration of the warning information based on the severity of the abnormal incident; the duration being positively correlated with the severity of the abnormal incident; and continuously sending the warning information to nearby vehicles during the duration.

[0013] Based on the aforementioned technical means, by determining the duration of the warning information in relation to the severity of the accident risk, and repeatedly pushing the warning information during the duration of the warning information, it is possible to ensure that all vehicles entering the target area can continuously receive effective warnings during the accident impact period (especially before the fault is repaired), avoiding the omission of dynamically moving vehicles due to a single push, preventing misoperation caused by the long-term retention of warning information, and reducing invalid transmission and resource consumption of the system.

[0014] Secondly, embodiments of this application provide an abnormal accident early warning system, including: a vehicle and a cloud server; the vehicle is configured to send abnormal accident information to the cloud server in response to detecting an abnormal accident; the cloud server is configured to perform abnormal accident early warning based on any of the optional abnormal accident early warning methods in the first aspect.

[0015] By receiving abnormal accident information reported by vehicles and combining core data such as the location, type, and images of the abnormal accidents, the target area is accurately determined, ensuring that the early warning focuses only on the actually affected area. Based on key accident information, the early warning is sent to nearby vehicles, achieving precise delivery of the warning information to the target vehicles. This avoids interference caused by irrelevant vehicles receiving the warning information and ensures that vehicles within the affected area receive the warning content in a timely manner. Compared to existing technologies that rely on manual labor and roadside fixed equipment to obtain abnormal accident information, using vehicles as sensing devices to directly report information eliminates the need for manual patrols or long-term monitoring by roadside equipment, improving the timeliness and coverage of accident information acquisition.

[0016] In some embodiments, the vehicle is further configured to, in response to receiving an abnormal accident reporting instruction, acquire surrounding environment image data; determine whether an abnormal accident has occurred based on the surrounding environment image data; the abnormal accident reporting instruction is triggered by at least one of the following methods: receiving an operation from a user to trigger a reporting physical button or a reporting virtual control; receiving a user's voice reporting accident information.

[0017] Because natural disasters such as earthquakes, floods, and landslides do not have fixed signals, vehicle-mounted sensors have difficulty automatically capturing and identifying them. However, manual triggering methods (physical buttons, virtual control operations, or voice commands) allow drivers to proactively initiate the image acquisition process as soon as they detect an anomaly. The vehicle then completes the accident determination and reporting based on the acquired surrounding environmental image data. This solves the limitation of automatic identification relying on fixed signals, and the standardized image acquisition and determination process ensures the effectiveness of the reported information. This allows natural disaster-related accidents that are originally difficult to capture to quickly enter the early warning system, improving the timeliness of abnormal accident reporting.

[0018] In some embodiments, the vehicle is also configured to send warning information to nearby vehicles based on a near-field communication module.

[0019] Because natural disasters have a wide impact, cloud servers must go through multiple stages, including data parsing and cross-node transmission, when distributing warnings in a targeted manner. This inevitably leads to network latency. Furthermore, nearby vehicles are close to the accident site, and their evacuation decision-making window is extremely short; warning delays can affect the effectiveness of evacuation efforts. Vehicles can send warning information directly to surrounding vehicles via near-field communication modules (such as V2X technology), eliminating the need for cloud relay and achieving second-level or even sub-second delivery of warning information, minimizing response time for nearby vehicles.

[0020] Thirdly, this application provides an abnormal accident early warning device, comprising: a determining unit, configured to, in response to receiving abnormal accident information sent by a vehicle, determine a target area affected by the abnormal accident based on the location and type of the abnormal accident indicated in the abnormal accident information; the abnormal accident information includes: the type of the abnormal accident, the location of the abnormal accident, and image data of the abnormal accident; the accident type includes at least one of the following: road collapse, bridge deck fracture, flood, landslide, and rockfall; a generating unit, configured to generate early warning information based on the type of the abnormal accident, the location of the abnormal accident, and the image data of the abnormal accident; the early warning information is used to prompt other vehicles passing through the target area to take evasive action; and a sending unit, configured to send the early warning information to nearby vehicles, wherein nearby vehicles are vehicles whose distance from the target area is less than or equal to a preset distance range.

[0021] Fourthly, this application provides an electronic device, including: a processor and a memory configured to store processor-executable instructions; wherein the processor is configured to execute the instructions to implement any of the optional abnormal accident early warning methods in the first aspect described above.

[0022] Fifthly, a computer-readable storage medium is provided, on which instructions are stored, such that when the instructions in the computer-readable storage medium are executed by a device, the device is able to perform any of the optional abnormal accident early warning methods described in the first aspect.

[0023] In a sixth aspect, a computer program product is provided, the computer program product including computer instructions that, when executed on a processor of a device, enable the device to perform any of the optional abnormal incident warning methods described in the first aspect above. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the background art, the accompanying drawings used in the embodiments of this application will be described below.

[0025] Figure 1 This is a schematic diagram of the structure of an abnormal accident early warning system disclosed in an embodiment of this application; Figure 2 This is a communication architecture diagram of the vehicle and cloud server disclosed in an embodiment of this application; Figure 3 This is a flowchart illustrating the process of a company response center handling abnormal incident information, as disclosed in an embodiment of this application. Figure 4 This is a topic hierarchy diagram of message sending via the MQTT protocol disclosed in the embodiments of this application; Figure 5 This is a schematic diagram of the cloud server's process for handling abnormal incident information as disclosed in an embodiment of this application; Figure 6 This is a schematic diagram illustrating the interaction between the cloud server and surrounding vehicles as disclosed in an embodiment of this application. Figure 7 This is a schematic diagram of the process for reporting abnormal accident information by a vehicle, as disclosed in an embodiment of this application. Figure 8 This is a flowchart illustrating an abnormal accident early warning method disclosed in an embodiment of this application; Figure 9 This is a flowchart illustrating another abnormal accident early warning method disclosed in an embodiment of this application; Figure 10 This is a schematic diagram of the structure of an abnormal accident early warning device disclosed in an embodiment of this application; Figure 11 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of this application. Detailed Implementation

[0026] The terms "first," "second," etc., are used for descriptive purposes only and have no sequential or technical meaning, nor should they be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Directional terms used in this application, such as "upper," "lower," "front," "rear," "left," "right," "inner," and "outer," are merely for reference to the orientation shown in the accompanying drawings. The use of directional terms is for better and clearer explanation and understanding of this application, and does not indicate the orientation of the referred device or component in an actual application scenario.

[0027] The embodiments of this application are described below with reference to the accompanying drawings.

[0028] Please see Figure 1 , Figure 1 This is a schematic diagram of the abnormal accident early warning system disclosed in an embodiment of this application. The abnormal accident early warning system of this application includes: a vehicle 101 and a cloud server 102.

[0029] In some embodiments, vehicle 101 is configured to send abnormal incident information to cloud server 102 in response to detecting an abnormal incident.

[0030] Among them, the vehicle can be, but is not limited to, any type of vehicle that has the ability to communicate with the cloud server, such as a pure electric vehicle (PEV / BEV), a hybrid electric vehicle (HEV), a range-extended electric vehicle (REEV), a plug-in hybrid electric vehicle (PHEV), or a new energy vehicle.

[0031] In the embodiments of this application, such as Figure 2 As shown, the vehicle includes an onboard perception unit, an onboard main control unit, and a communication unit, enabling it to send abnormal incident information to a cloud server when an abnormal incident is detected. The onboard perception unit includes cameras, millimeter-wave radar, a GPS positioning unit, and vehicle attitude sensors, used to collect image data, surrounding environment data, location data, and vehicle operating status data related to the abnormal incident. The onboard main control unit performs keyframe extraction and compression processing on the collected image data to generate a low-resolution preview image and metadata (including shooting timestamps and location information). The communication unit uses message queuing telemetry transport (MQTT) or vehicle-to-everything (V2X) technology to achieve information transmission with the cloud server.

[0032] See again Figure 2 Vehicles and cloud servers can communicate based on the MQTT protocol. The communication process involves the vehicle reporting abnormal incident information to the cloud MQTT message broker node. The cloud MQTT broker node forwards this information to the company's response center. The company's response center analyzes the abnormal incident information, obtains the analysis results, and returns them to the cloud MQTT broker node. These results include early warning information. Based on the incident type and location in the abnormal incident information, the company's response center determines the target area and generates early warning information based on the incident type, location, and image data. After receiving the analysis results, the cloud MQTT broker node sends early warning information to surrounding vehicles (nearby vehicles) and synchronizes the abnormal incident information and warning information with the National Transportation Center.

[0033] In some embodiments, such as Figure 3As shown, the company's response center receives abnormal incident information sent by the broker. First, it verifies the message source and signature integrity, filtering out abnormal or duplicate reports. Then, it calls a deep learning model (fusing image, GPS, speed, attitude, etc.) to identify the incident type and severity, obtaining the identification result. If it is determined to be a common collision (or minor anomaly), it is only archived in the database, and a manual follow-up is prompted. If it is identified as a road anomaly such as road collapse, bridge collapse, rockfall, or flood, a warning data packet is automatically generated.

[0034] The company's response center calculates the affected area using a geographic information system and generates a topic located by kilometer marker, such as / traffic / alert / highway / G4 / 195-205km, which indicates that an abnormal traffic accident has occurred on the section of G4 highway from kilometer 195 to kilometer 205. This topic is a dedicated warning information release topic for the corresponding road section. After the alarm is generated, it is pushed back to the MQTT Broker node, and the Broker releases it to nearby vehicles based on the target area.

[0035] like Figure 4 As shown, the topic hierarchy for sending messages via the MQTT protocol is divided into an event reporting branch on the left and a traffic warning branch on the right.

[0036] The left branch is rooted at / incident / , with sub-nodes / report / and / disaster / . / report / represents the reporting path for routine vehicle events (such as vehicle collisions, traffic jams, and other non-natural disaster incidents), while / disaster / represents the reporting path for natural disaster incidents (such as landslides, floods, rockfalls, and other sudden incidents caused by natural factors). / report / is further associated with the vehicle identification number path {VIN}, which is used for reporting specific event information.

[0037] The right-hand branch takes / traffic / as its root node and refines it layer by layer through the / alert / child nodes to the geographical levels such as province, city, and {road_id} (road name), so as to accurately locate the scope of the warning information release according to the administrative region and road number.

[0038] In some embodiments, the company's response center may also determine the synchronization path based on the alarm level (Level I to Level III): Alarm Level I: internal company response and regional vehicle warning only; Alarm Level II: synchronization with the national transportation center; Alarm Level III: triggering the national emergency broadcast system interface.

[0039] In some embodiments, cloud server 102 is configured to provide anomaly warnings based on the anomaly warning method described in the following embodiments.

[0040] The cloud server can be a distributed cloud server cluster, including data receiving nodes, early warning processing nodes, MQTT Broker nodes and distribution nodes, with high concurrency processing capabilities and low latency response characteristics, which can support the information reporting and early warning information distribution of massive vehicles.

[0041] In this embodiment, the cloud server receives abnormal accident information reported by vehicles through data receiving nodes. The early warning processing node verifies the authenticity of the accident, determines the risk level, and identifies the target area based on preset rules. The early warning information is then published to the corresponding geographic subscription channel through the MQTT Broker node. Finally, the distribution node accurately pushes the early warning information based on the vehicle subscription relationship, ensuring that vehicles in the affected area receive the early warning information in a timely manner.

[0042] like Figure 5 As shown, a multi-node MQTT Broker cluster is deployed on the cloud server. Message synchronization between nodes is achieved through a consistency protocol (such as Raft), ensuring high availability and low latency.

[0043] After receiving abnormal accident information reported by vehicles, the cloud server verifies the validity of the signature and certificate. It stores the information in a message queue and forwards it to the company's response center for further intelligent analysis, awaiting the return of the analysis results. If the MQTT Broker node detects duplicate reports or multiple abnormal reports from the same road segment, it can trigger a rapid aggregation mechanism, marking the event on that road segment as a "high-priority anomaly" and generating a draft early warning in advance.

[0044] After the company's response center generates an alert, Broker generates a topic based on geographical hierarchy (e.g., / traffic / alert / G4 / 190-210km) and publishes the alert message. The `retain` attribute and message time-to-live (TTL) are set for each message to ensure that vehicles entering the affected road segment within the alert's validity period can still automatically receive the information. After the TTL expires, Broker automatically clears expired messages to prevent false alarms from persisting for extended periods. Finally, the alert information is pushed to nearby vehicles and the National Traffic Center.

[0045] In this embodiment, the Broker publishes messages in a hierarchical format based on traffic domain / alert type / road sign / road segment range (e.g., / traffic / alert / highway / G4 / 195-205km in the example). This format corresponds exactly to the geographic range subscription channels pre-subscribed by vehicles. While driving, vehicles dynamically subscribe to the hierarchical topic of the current road segment. The Broker node only distributes messages to vehicles that have subscribed to the corresponding hierarchical topic, avoiding pushing messages to vehicles on irrelevant road segments.

[0046] The `retain` property in the MQTT protocol allows the broker to retain the latest message under a topic, rather than only pushing it to subscribers who were online at the time of publication. Specifically, when an alert is published, vehicles already within the 190-210km range and subscribed to the corresponding topic will receive the alert in real time. If a new vehicle subsequently enters this range (within the message's lifespan), the vehicle will automatically subscribe to that range's topic. At this point, the broker will automatically push the latest retained alert message to that vehicle, without requiring the response center to republish it, ensuring that no vehicle entering the range within the alert's validity period misses the warning.

[0047] Regarding message lifespan, Broker sets a fixed validity period for each alert message. Within the validity period, the message is retained and can be distributed to subscribed vehicles. If the TTL is exceeded, Broker will automatically delete the message and will no longer push it to newly subscribed vehicles.

[0048] It should be noted that, in addition to MQTT, the Broker supports interconnection with edge computing nodes to enable rapid distribution of alerts near road sections. For example, edge nodes can send alerts to vehicles without internet access via 5G C-V2X, RSU (Roadside Unit), or satellite relay links. The Broker also synchronizes high-level alert information to a dedicated MQTT interface at the National Transportation Center.

[0049] In some embodiments, the client (nearby vehicle) subscribes to the MQTT topic of the corresponding road segment in real time, such as / traffic / alert / highway / G4 / 195-205km. When a retain message or a new alarm is detected, it immediately responds with a warning message and provides a prompt, such as a pop-up window on the vehicle terminal or a voice broadcast reminder, while simultaneously triggering automatic navigation route planning.

[0050] In some embodiments, such as Figure 6As shown, after nearby vehicles receive the warning information sent by the cloud-based Broker node, the central control system prompts the driver through display and voice broadcast. For example, a pop-up window displays the accident type, distance, and warning level, while simultaneously broadcasting the warning. If the driver does not respond to the prompt, the system can escalate to flashing warning lights and vibration feedback.

[0051] See again Figure 6 If the vehicle has autonomous driving capabilities, the system will automatically adjust its behavior based on the type of warning: if the abnormal incident is a road blockage, it will reduce speed and plan an alternative route; if the abnormal incident is a bridge collapse, it will prohibit entry and apply the brakes; if the abnormal incident is a flood / rockfall, it will trigger route replanning and risk avoidance.

[0052] In some embodiments, V2X near-field communication technology can be used in conjunction with a cloud-based MQTT Broker to achieve a closed-loop linkage between vehicles, the cloud, and other networks. If nearby vehicles detect a similar alarm, it can be directly forwarded to other vehicles not connected to the network via V2X, achieving self-organized propagation within the area (Mesh Relay). The cloud-based Broker is responsible for continuously maintaining vehicle status and message validity, forming a complete detection-analysis-broadcast-risk avoidance closed loop.

[0053] Therefore, by receiving abnormal accident information reported by vehicles and combining core data such as the location, type, and images of the abnormal accidents, the target area is accurately determined, ensuring that the early warning focuses only on the actually affected area. Based on key accident information, the early warning is sent to nearby vehicles, achieving precise delivery of the warning information to the target vehicles. This avoids interference caused by irrelevant vehicles receiving the warning information and ensures that vehicles within the affected area receive the warning content in a timely manner. Compared to existing technologies that obtain abnormal accident information through manual labor and fixed roadside equipment, using vehicles as sensing devices to directly report information eliminates the need for manual patrols or long-term monitoring by roadside equipment, improving the timeliness and coverage of accident information acquisition, thereby shortening the early warning time for abnormal accidents.

[0054] In some embodiments, since natural disasters do not have fixed collision signals and cannot be automatically triggered for identification by traditional sensors, judgment must be made in conjunction with image data. To improve the timeliness and coverage of natural disaster identification, a manual triggering method can be set to actively initiate the image acquisition and judgment process, compensating for the limitations of automatic identification. Therefore, the vehicle is also configured to acquire surrounding environmental image data in response to receiving an abnormal accident reporting command; and to determine whether an abnormal accident has occurred based on the surrounding environmental image data.

[0055] The abnormal incident reporting command is triggered in at least one of the following ways: receiving the user's operation of triggering the reporting physical button or triggering the reporting virtual control; receiving the user's voice report of incident information.

[0056] In this embodiment, when a user-triggered abnormal accident reporting command is received, the vehicle acquires surrounding environmental image data by using onboard front-view and side-view cameras to collect images of the suspected accident area. Then, the vehicle's image recognition module calls a preset natural disaster feature library (including contour and texture feature templates of typical accidents such as road collapses, rockfalls, floods, and landslides). The collected image data is compared with the preset natural disaster feature library using a feature point matching algorithm. If the matching similarity exceeds a preset threshold, it is determined to be an abnormal accident, and the accident type corresponding to the matched feature template is determined as the accident type of the abnormal accident.

[0057] For example, when a driver discovers falling rocks on the road ahead, pressing the physical accident reporting button on the vehicle's central control panel triggers a reporting command. The vehicle immediately takes continuous pictures of the road covered by falling rocks using the vehicle's camera. The image recognition module compares the captured images with falling rock templates in the feature library. If an abnormal accident is determined to have occurred, the subsequent information reporting process is automatically initiated.

[0058] like Figure 7 As shown, the vehicle is equipped with a one-click reporting button. Users can generate a button signal (an abnormal accident reporting command) by triggering the one-click reporting button (located on the central control screen, steering wheel, or voice command entry). The button signal is transmitted to the vehicle's main control unit via the CAN bus. In addition to manual triggering, the vehicle's main control unit also monitors vehicle attitude changes and surrounding environmental data in real time through multiple sensors (including accelerometers, gyroscopes, wheel speed sensors, radar, and cameras).

[0059] When a collision, sudden deceleration, or abnormal posture change is detected (such as the vehicle suddenly tilting, hovering, or the ground collapsing), the vehicle automatically triggers the reporting process to achieve unattended emergency reporting.

[0060] Upon triggering the abnormal incident reporting command, the system controls the cameras to collect images from the front, rear, left, and right. Combined with the vehicle's inertial sensors and global navigation system module, it records abnormal incident information, including: on-site images, video clips, GPS coordinates, vehicle speed, timestamps, heading angles, and other data. The onboard main control unit incorporates a fast visual feature extraction algorithm to perform preliminary local classification of abnormal scenes, distinguishing between types such as "collision," "road collapse," "obstacle blockage," and "flood / rockfall."

[0061] To improve reporting speed, image data is extracted and compressed locally, with only keyframes and low-resolution previews uploaded. Data packets are encrypted with AES-256 and authenticated via MQTT certificates. Abnormal incident information is encapsulated as a JSON data packet and uploaded to the cloud broker node via MQTT+TLS.

[0062] It should be understood that since natural disasters such as earthquakes, floods, and landslides do not have fixed signals, vehicle-mounted sensors have difficulty automatically capturing and identifying them. However, manual triggering methods (physical buttons, virtual control operations, or voice commands) allow drivers to proactively initiate the image acquisition process as soon as they detect an anomaly. The vehicle then completes the accident determination and reporting based on the acquired surrounding environmental image data. This solves the limitation of automatic identification relying on fixed signals, and the standardized image acquisition and determination process ensures the effectiveness of the reported information. This allows natural disaster-related accidents that are originally difficult to capture to quickly enter the early warning system, improving the timeliness of abnormal accident reporting.

[0063] In some embodiments, due to the wide-ranging impact of natural disasters, the targeted distribution of early warning information by cloud servers requires data processing and subscription matching, resulting in network latency. Furthermore, nearby vehicles are close to the accident site and have limited time to evacuate, necessitating a faster early warning delivery method. Therefore, vehicles are also configured to send early warning information to nearby vehicles via a near-field communication module.

[0064] In this embodiment, the vehicle's near-field communication module adopts V2X communication technology, which supports direct data transmission between vehicles (V2V). When the vehicle determines that an abnormal accident has occurred, the near-field communication module is automatically activated and obtains other vehicles in the target area that can establish a communication connection with the vehicle. After establishing communication, it sends warning information to these vehicles, thereby achieving the desired result.

[0065] For example, after vehicle A determines that there is a road collapse at 195km on the G4 expressway, its V2X near-field communication module is immediately activated. Based on its own GPS positioning, it determines the target area and then sends a warning message to vehicles B and C in the target area. The message contains the core content that there is a road collapse 195km ahead and suggests slowing down, avoiding and detouring. After receiving the message, vehicles B and C simultaneously trigger a voice broadcast to remind the driver.

[0066] Understandably, due to the wide-ranging impact of natural disasters, cloud servers must go through multiple stages, including data parsing and cross-node transmission, when distributing warnings in a targeted manner. This inevitably leads to network latency. Furthermore, nearby vehicles are close to the accident site, and their evacuation decision-making window is extremely short; warning delays may affect the effectiveness of evacuation efforts. Vehicles can directly send warning information to surrounding vehicles via near-field communication modules (such as V2X technology), eliminating the need for cloud relay and achieving second-level or even sub-second delivery of warning information, minimizing the response time for nearby vehicles.

[0067] Please see Figure 8 Figure 8 is a flowchart illustrating the abnormal accident early warning method disclosed in this application, which can be applied to the cloud server in the above embodiments and includes the following steps: S801. In response to receiving abnormal accident information sent by the vehicle, determine the target area affected by the abnormal accident based on the location and type of the abnormal accident indicated in the abnormal accident information.

[0068] The abnormal accident information includes: the type of abnormal accident, the location of the abnormal accident, and image data of the abnormal accident. The accident type includes at least one of the following: road collapse, bridge deck fracture, flood, landslide, and rockfall.

[0069] As one possible implementation, a pre-defined area expansion rule is established based on the accident type: road collapse / bridge breakage extends outward by 50-100 meters, flood / landslide by 100-200 meters, and rockfall by 30-50 meters. An initial impact range is formed based on the location data of the abnormal accident (such as GPS coordinates and road kilometer markers) combined with the accident type. Spatial distribution features (such as flood coverage area, rockfall radius, and road damage area) are extracted from the abnormal accident images and converted into actual physical dimensions through pixel calibration. The initial impact range is then adjusted. If the image shows that the actual flood coverage area is larger than the pre-defined expansion value, the target area boundary is expanded according to the actual size. The boundary of the target area is further adjusted based on the number of road lanes and speed limits. If the number of road lanes is small and the speed is relatively high, the area of ​​the target area needs to be increased. For example, the accident target area on a highway needs to be extended an additional 50 meters upstream and downstream to ensure coverage of the safe distance required for vehicle braking and avoidance.

[0070] It should be noted that, in addition to the accident types mentioned above, abnormal accident information can also be identified in cases of vehicle collisions and traffic congestion, further expanding the coverage of the early warning system and improving the comprehensiveness of road traffic safety. Specifically, abnormal accident information in vehicle collision scenarios can include collision type (e.g., frontal collision, side collision), collision severity (determined based on vehicle airbag deployment status and acceleration sensor data), collision location, and on-site images. Abnormal accident information in traffic congestion scenarios can include the extent of the congestion, the start time of the congestion, average traffic speed, and image data of the congested area.

[0071] In this embodiment of the application, for vehicle collision scenarios, the vehicle can automatically trigger the collection of abnormal accident information through the airbag controller and the vehicle body acceleration sensor without manual operation, ensuring that the information is reported quickly after the collision. For traffic congestion scenarios, the vehicle can combine the road traffic flow images collected by the vehicle camera (identifying lane occupancy rate and vehicle queue length) with the real-time vehicle speed data to automatically determine the congestion level and generate abnormal accident information.

[0072] It should be understood that in the event of an abnormal accident, vehicles transmit abnormal accident information. As the main sensing entity of the abnormal accident, especially in remote road sections or monitoring blind spots, vehicles can capture and report key information immediately, preventing the impact of the accident from escalating due to information delays. At the same time, the information transmitted by vehicles undergoes local preprocessing and preliminary verification to ensure the authenticity and validity of the data. It can quickly transmit the real-time situation of the accident scene (such as location, type, images, and other core data) to the cloud server, providing key data support for the cloud to quickly initiate early warning processes and accurately formulate risk avoidance strategies.

[0073] S802. Generate early warning information based on the type of abnormal accident, the location of the abnormal accident, and the image data of the abnormal accident.

[0074] The warning information is used to alert other vehicles passing through the target area to take evasive action.

[0075] In some embodiments, the target area is determined by: extracting spatial distribution features of the abnormal accidents from the image data of the abnormal accidents in the abnormal accident information; the spatial distribution features include at least one of the following: the scope of the accident's impact and the degree of road surface damage; determining the basic impact area of ​​the abnormal accident based on the scope of the accident's impact and the accident type of the abnormal accident; determining the degree of risk of the abnormal accident causing potential accidents based on the degree of road surface damage; and correcting the basic impact area based on the degree of risk of causing potential accidents to obtain the target area.

[0076] As one possible approach, spatial distribution features are extracted from abnormal accident image data using image segmentation and edge detection algorithms to obtain the accident's impact range and road surface damage degree. Accident impact range: quantifies the actual physical area covered by the accident (e.g., the area of ​​road surface submerged by floods, the lateral / longitudinal distance of scattered rocks, the number of lanes occupied by landslides). Road surface damage degree: identifies damage types (cracks, potholes, collapses) and quantifies them using pixel comparison (e.g., crack length / width, collapse depth / area), classifying them into three levels: minor, moderate, and severe.

[0077] Using the accident impact range extracted from the image as the initial boundary (e.g., the rockfall area is 5 meters horizontally and 20 meters vertically), and combining it with the preset expansion amount corresponding to the accident type, the basic impact area is obtained. Different types of accidents correspond to different expansion amounts: road collapses expand outward by 30-50 meters based on the actual impact range, rockfalls expand by 10-20 meters based on the scattered area, and floods expand by 50-100 meters based on the inundation area, thus forming the basic impact area.

[0078] Potential risks include secondary collapses, road surface cracking and expansion, and debris splashing, which may lead to subsequent accidents. Based on the pre-defined correspondence between the degree of road damage and potential risks, the risk level of potential accidents caused by abnormal accidents is obtained: minor damage corresponds to low risk, moderate damage corresponds to medium risk, and severe damage corresponds to high risk.

[0079] The corresponding regional correction rules are determined based on the potential risk level of the abnormal accident. The higher the risk level, the greater the expansion of the basic impact area, and the two are positively correlated. When correcting the basic impact area, a fixed-direction expansion method can be adopted, prioritizing the extension towards areas with severe road damage to ensure that the warning range accurately covers high-risk areas.

[0080] It should be understood that extracting two major spatial features from abnormal accident images—the scope of the accident's impact and the degree of road surface damage—first determines the basic impact area based on the scope of impact and the type of accident, then predicts potential accident risks based on the degree of road surface damage, and finally corrects the basic area according to the risk level. This not only considers the direct impact of the current accident but also predicts potential risks based on the degree of road surface damage, thereby improving the overall accuracy of the target area.

[0081] S803: Send the warning information to nearby vehicles.

[0082] Among them, nearby vehicles are those whose distance from the target area is less than or equal to a preset distance range.

[0083] In some embodiments, sending warning information to nearby vehicles includes: determining a target geographic range subscription channel corresponding to the abnormal accident based on the location of the abnormal accident; the geographic range subscription channel is a message transmission channel for sending warning information; each geographic range subscription channel corresponds to a geographic range of a fixed road segment; the warning information is published to the target geographic range subscription channel; the target geographic range subscription channel is used to send warning information to nearby vehicles that have subscribed to the target geographic range subscription channel.

[0084] One possible implementation involves identifying the target geographic area subscription channel based on the location of the abnormal accident (e.g., road kilometer markers, GPS coordinates) (e.g., if the accident is at kilometer 195 on the G4 expressway, the matching channel would be / traffic / alert / G4 / 190-200km). The MQTT Broker node uploads the warning information to the target geographic area subscription channel, which automatically associates with all nearby vehicles subscribed to that channel. Warning information is then synchronously sent to subscribed vehicles via the target geographic area subscription channel; vehicles not subscribed to that channel will not receive the warning, achieving precise push notifications to vehicles within the affected road segment.

[0085] It should be understood that during vehicle operation, the system dynamically subscribes to channels corresponding to the geographical area of ​​the current road segment based on real-time location. This eliminates the need for continuous location reporting to the cloud to request alerts, significantly reducing the communication load between vehicles and the cloud. Furthermore, the Broker node only needs to publish the alert information to the corresponding channel to automatically reach all subscribed vehicles, avoiding the complex calculations required for location filtering across the entire network and improving alert distribution efficiency. Simultaneously, the strong binding between channels and fixed road areas fundamentally prevents the invalid push of alert information to vehicles in unaffected areas, reducing information interference to onboard terminals and lowering the transmission pressure on cloud servers.

[0086] In some embodiments, sending warning information to nearby vehicles includes: determining the duration of the warning information based on the severity of the abnormal incident; the duration being positively correlated with the severity of the abnormal incident; and continuously sending the warning information to nearby vehicles during the duration.

[0087] One possible implementation is to determine the duration of the warning message based on a preset rule corresponding to the severity and duration of the accident, as well as the estimated time for fault repair. Within the set duration, the system repeatedly sends the warning message to nearby vehicles at a fixed frequency (e.g., once every 30 seconds), or continuously broadcasts it through a geographic subscription channel. After the duration expires, the system stops pushing messages to avoid expired warnings leading to erroneous actions.

[0088] It should be understood that by determining the duration of the warning information in relation to the severity of the accident risk, and repeatedly pushing the warning information during the duration of the warning information, it is possible to ensure that all vehicles entering the target area can continuously receive effective warnings during the accident impact period (especially before the fault is repaired), avoiding the omission of dynamically moving vehicles due to a single push, preventing misoperation caused by the long-term retention of warning information, and reducing invalid transmission and resource consumption of the system.

[0089] Therefore, by receiving abnormal accident information reported by vehicles and combining core data such as the location, type, and images of the abnormal accidents, the target area is accurately determined, ensuring that the early warning focuses only on the actually affected area. Based on key accident information, the early warning is sent to nearby vehicles, achieving precise delivery of the warning information to the target vehicles. This avoids interference caused by irrelevant vehicles receiving the warning information and ensures that vehicles within the affected area receive the warning content in a timely manner. Compared to existing technologies that obtain abnormal accident information through manual labor and roadside fixed equipment, using vehicles as sensing devices to directly report information eliminates the need for manual patrols or long-term monitoring by roadside equipment, improving the timeliness and coverage of accident information acquisition.

[0090] In some embodiments, the abnormal accident information further includes: vehicle attitude information. Combining the vehicle attitude information can accurately determine the severity of the accident, thereby improving the accuracy of the target area. Therefore, such as Figure 9 As shown, the above S802 includes the following steps: S901. Taking the location of the abnormal accident as the center, the area within a preset distance from the abnormal accident is determined as the initial area of ​​the abnormal accident.

[0091] For example, a circular or rectangular area along the road is delineated at a preset distance (e.g., 50 meters by default) centered on the location of the abnormal accident (e.g., GPS coordinates, road kilometer markers) as the initial area.

[0092] S902. Input the vehicle's attitude information into the abnormal accident severity determination model. The abnormal accident severity determination model determines the degree to which the vehicle deviates from its normal driving posture based on the vehicle's attitude information, and thus determines the severity of the abnormal accident.

[0093] As one possible implementation, vehicle attitude information includes data such as vehicle tilt angle, acceleration, steering angle, and rollover risk value. This vehicle attitude information is input into a pre-defined abnormal accident severity determination model. The model calculates the degree of deviation between the actual vehicle attitude and the normal driving attitude by comparing it to thresholds for normal driving attitude (e.g., tilt angle ≤ 5°, steering angle ≤ 15°). The severity level is then classified according to the degree of deviation.

[0094] The pre-defined model for determining the severity of abnormal accidents is trained based on historical vehicle driving posture sample data. The training data covers normal driving posture data under different vehicle models and road conditions, as well as posture data corresponding to various abnormal accident scenarios (such as collisions, rollovers, and bumps). By labeling the posture parameters in the sample data with the actual severity level of the accident, the pre-defined model for determining the severity of abnormal accidents is obtained after iterative optimization. After inputting real-time vehicle posture information, the deviation quantification value between the actual posture and the normal posture threshold is calculated using a pre-defined algorithm (such as deviation coefficient calculation, weighted scoring method, etc.). Then, the corresponding severity level is mapped according to the deviation value range (e.g., deviation value less than 20% is mild, greater than or equal to 20% and less than or equal to 50% is moderate, and greater than 50% is severe).

[0095] S903. Determine the impact of abnormal accidents based on their accident types.

[0096] One possible approach is to match the accident type with a preset impact level to obtain the impact level of the current abnormal accident. For example, road collapse and bridge breakage correspond to "high impact" (wide impact range and long duration); rockfall and small landslide correspond to "medium impact"; and minor water accumulation and sporadic rockfall correspond to "low impact".

[0097] S904. Determine the regional adjustment change amount based on the severity and impact of the abnormal accident.

[0098] Among them, the amount of regional adjustment change is positively correlated with the severity; the amount of regional adjustment change is positively correlated with the degree of impact.

[0099] As one possible implementation, a mapping table is pre-defined to correspond regional adjustment amounts to severity and impact levels. Mild severity corresponds to a base adjustment amount, moderate severity to a medium adjustment amount, and severe severity to a high adjustment amount. An impact coefficient is determined based on the degree of impact. For example, high, medium, and low impact levels correspond to 1.5 times, 1 time, and 0.5 times the base adjustment amount, respectively. The regional adjustment change is the product of the severity-related adjustment amount and the impact coefficient.

[0100] S905. Based on the regional adjustment change, adjust the initial region to obtain the target region.

[0101] It should be understood that the initial area is determined based on the accident location, the severity of the accident is determined by the vehicle posture information, the degree of impact is determined by the accident type, and the area range is adjusted positively according to the severity and degree of impact to finally obtain a target area that fits the actual risk, so as to avoid the early warning failing due to the area being too large or too small.

[0102] The above primarily describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the above functions, the abnormal accident early warning device or electronic device includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0103] This application embodiment can, based on the above method, exemplarily divide an abnormal accident early warning device or electronic device into functional modules. For example, the abnormal accident early warning device or electronic device may include functional modules corresponding to each functional division, or two or more functions may be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division; in actual implementation, there may be other division methods.

[0104] Please see Figure 10 , Figure 10 This is a schematic diagram of the structure of the abnormal accident early warning device disclosed in the embodiments of this application. The abnormal accident early warning device 1000 includes: a determination unit 1010, a generation unit 1020, and a sending unit 1030.

[0105] The determining unit 1010 is configured to, in response to receiving abnormal accident information sent by a vehicle, determine the target area affected by the abnormal accident based on the location of the abnormal accident indicated in the abnormal accident information and the accident type of the abnormal accident; the abnormal accident information includes: the accident type of the abnormal accident, the location of the abnormal accident, and image data of the abnormal accident; the accident type includes at least one of the following: road collapse, bridge deck fracture, flood, landslide, and rockfall.

[0106] The generation unit 1020 is used to generate early warning information based on the type of the abnormal accident, the location of the abnormal accident, and the image data of the abnormal accident; the early warning information is used to prompt other vehicles passing through the target area to take evasive action.

[0107] The sending unit 1030 is used to send the warning information to nearby vehicles, wherein the nearby vehicles are vehicles whose distance from the target area is less than or equal to a preset distance range.

[0108] In some embodiments, the abnormal accident information further includes: vehicle posture information; a determination unit 1010, specifically configured to determine an area within a preset distance from the abnormal accident as the initial area of ​​the abnormal accident, centered on the location of the abnormal accident; input the vehicle posture information into an abnormal accident severity determination model, and determine the degree to which the vehicle deviates from its normal driving posture based on the vehicle posture information, thereby determining the severity of the abnormal accident; determine the degree of impact of the abnormal accident based on the accident type; determine the area adjustment change amount based on the severity and impact of the abnormal accident; the area adjustment change amount is positively correlated with the severity; the area adjustment change amount is positively correlated with the degree of impact; and adjust the initial area based on the area adjustment change amount to obtain the target area.

[0109] In some embodiments, the determining unit 1010 is specifically used to extract spatial distribution features of the abnormal accident in the image based on the image data of the abnormal accident in the abnormal accident information; the spatial distribution features include at least one of the following: the scope of the accident's impact and the degree of road surface damage; based on the scope of the accident's impact and the accident type of the abnormal accident, determine the basic impact area of ​​the abnormal accident; based on the degree of road surface damage, determine the risk level of the abnormal accident causing a potential accident; based on the risk level of causing a potential accident, correct the basic impact area to obtain the target area.

[0110] In some embodiments, the sending unit 1030 is specifically used to determine the target geographic range subscription channel corresponding to the abnormal accident based on the location of the abnormal accident; the geographic range subscription channel is used as a message transmission channel for sending early warning information; each geographic range subscription channel corresponds to a geographic range of a fixed road segment; the early warning information is published to the target geographic range subscription channel; the target geographic range subscription channel is used to send early warning information to nearby vehicles that have subscribed to the target geographic range subscription channel.

[0111] In some embodiments, the sending unit 1030 is specifically configured to determine the duration of the warning information based on the severity of the abnormal accident; the duration is positively correlated with the severity of the abnormal accident; and the warning information is continuously sent to nearby vehicles during the duration.

[0112] Please see Figure 11 The electronic device 1100 provided in this application embodiment includes, but is not limited to, a processor 1101 and a memory 1102.

[0113] The memory 1102 described above is used to store the executable instructions of the processor 1101. It is understood that the processor 1101 is configured to execute instructions to implement the abnormal incident warning method in the above embodiments.

[0114] It should be noted that those skilled in the art will understand that Figure 11 The electronic device structure shown does not constitute a limitation on electronic device 1100; electronic devices may include, but are not limited to, those described above. Figure 11 This may indicate more or fewer components, or combinations of certain components, or different component arrangements.

[0115] Processor 1101 is the control center of electronic device 1100. It connects various parts of the electronic device via various interfaces and lines. By running or executing software programs and / or modules stored in memory 1102, and by calling data stored in memory 1102, it performs various functions and processes data of electronic device 1100, thereby providing overall monitoring of electronic device 1100. Processor 1101 may include one or more processing units. Optionally, processor 1101 may integrate an application processor and a modem processor. The application processor mainly handles the operating system, user interface, and applications, while the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into processor 1101.

[0116] The memory 1102 can be used to store software programs and various data. The memory 1102 may primarily include a program storage area and a data storage area. The program storage area may store the operating system, application programs required by at least one functional module (such as a determination unit, processing unit, etc.), etc. Furthermore, the memory 1102 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0117] In an exemplary embodiment, a vehicle is also provided, including the aforementioned electronic equipment.

[0118] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 1102 including instructions, which can be executed by a processor 1101 of an electronic device 1100 to implement the methods in the above embodiments.

[0119] In actual implementation, Figure 10 The functions of each module can be provided by Figure 11 The processor 1101 calls the computer program stored in the memory 1102 to implement the process. The specific execution process can be found in the description of the method section in the previous embodiment, and will not be repeated here.

[0120] Optionally, the computer-readable storage medium may be a non-transitory computer-readable storage medium, such as a read-only memory (ROM), random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.

[0121] In an exemplary embodiment, this application also provides a computer program product including one or more instructions, which can be executed by the processor 1101 of the electronic device 1100 to perform the methods described above.

[0122] It should be noted that when one or more instructions in the computer-readable storage medium or computer program product are executed by the processor of an electronic device, they implement the various processes of the above method embodiments and achieve the same technical effect as the above method. To avoid repetition, they will not be described again here.

[0123] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0124] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another apparatus, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0125] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0126] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0127] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, essentially, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0128] It should be understood that the application of this application is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims. Those skilled in the art can understand that implementing all or part of the processes of the above embodiments and making equivalent changes according to the claims of this application still fall within the scope of this application.

Claims

1. A method for early warning of abnormal accidents, characterized in that, The method includes: In response to receiving abnormal accident information sent by a vehicle, the target area affected by the abnormal accident is determined based on the location of the abnormal accident and the accident type indicated in the abnormal accident information; the abnormal accident information includes: the accident type of the abnormal accident, the location of the abnormal accident, and image data of the abnormal accident; the accident type includes at least one of the following: road collapse, bridge deck fracture, flood, landslide, rockfall. Based on the type of the abnormal accident, the location of the abnormal accident, and the image data of the abnormal accident, a warning message is generated; the warning message is used to prompt other vehicles passing through the target area to take evasive action. The warning information is sent to nearby vehicles, which are vehicles whose distance from the target area is less than or equal to a preset distance range.

2. The abnormal accident early warning method according to claim 1, characterized in that, The abnormal accident information also includes: vehicle attitude information; determining the target area affected by the abnormal accident based on the location of the abnormal accident indicated in the abnormal accident information and the accident type of the abnormal accident includes: Centered on the location of the abnormal accident, the area within a preset distance from the abnormal accident is defined as the initial area of ​​the abnormal accident. The vehicle's attitude information is input into the abnormal accident severity determination model. The abnormal accident severity determination model determines the degree to which the vehicle deviates from its normal driving posture based on the vehicle's attitude information, and thus determines the severity of the abnormal accident. Based on the accident type of the abnormal accident, determine the degree of impact of the abnormal accident; The regional adjustment change is determined based on the severity and impact of the abnormal incident; the regional adjustment change is positively correlated with the severity; the regional adjustment change is positively correlated with the impact. Based on the change in the region, the initial region is adjusted to obtain the target region.

3. The abnormal accident early warning method according to claim 1, characterized in that, The target area is determined in the following way: Based on the image data of the abnormal accidents in the abnormal accident information, the spatial distribution features of the abnormal accidents in the images are extracted; the spatial distribution features include at least one of the following: the scope of the accident's impact and the degree of road surface damage. Based on the scope of the accident's impact and the type of the abnormal accident, the basic impact area of ​​the abnormal accident is determined; Based on the extent of road surface damage, determine the degree of risk of the abnormal accident triggering a potential accident; Based on the risk level of the potential accident, the basic impact area is adjusted to obtain the target area.

4. The abnormal accident early warning method according to claim 1, characterized in that, Sending the warning information to nearby vehicles includes: Based on the location of the abnormal incident, a target geographic range subscription channel corresponding to the abnormal incident is determined; the geographic range subscription channel is used as a message transmission channel for sending the early warning information; each geographic range subscription channel corresponds to a geographic range of a fixed road segment. The warning information is published to the target geographic area subscription channel; the target geographic area subscription channel is used to send the warning information to nearby vehicles that have subscribed to the target geographic area subscription channel.

5. The abnormal accident early warning method according to claim 1, characterized in that, Sending the warning information to nearby vehicles includes: The duration of the warning message is determined based on the severity of the abnormal incident; the duration is positively correlated with the severity of the abnormal incident. During the specified duration, the warning information is continuously sent to the nearby vehicles.

6. An abnormal accident early warning system, characterized in that, include: Vehicles and cloud servers; The vehicle is configured to send abnormal incident information to the cloud server in response to the detection of an abnormal incident; The cloud server is configured to perform abnormal incident early warning based on the abnormal incident early warning method according to any one of claims 1-5.

7. The abnormal accident early warning system according to claim 6, characterized in that, The vehicle is also configured to, In response to receiving an abnormal incident reporting command, acquire surrounding environmental image data; Based on surrounding environmental image data, determine whether an abnormal accident has occurred; The abnormal incident reporting command is triggered in at least one of the following ways: Receive user-triggered reports of physical button presses or virtual control actions; The system received a voice message from a user reporting an incident.

8. The abnormal accident early warning system according to claim 6, characterized in that, The vehicle is also configured to, The warning information is sent to nearby vehicles based on the near-field communication module.

9. An abnormal accident early warning device, characterized in that, include: The determining unit is configured to, in response to receiving abnormal accident information sent by a vehicle, determine the target area affected by the abnormal accident based on the location of the abnormal accident indicated in the abnormal accident information and the accident type of the abnormal accident. The abnormal accident information includes: the accident type of the abnormal accident, the location of the abnormal accident, and the image data of the abnormal accident; the accident type includes at least one of the following: road collapse, bridge deck fracture, flood, landslide, and rockfall; The generation unit is used to generate early warning information based on the type of the abnormal accident, the location of the abnormal accident, and the image data of the abnormal accident; the early warning information is used to prompt other vehicles passing through the target area to take evasive action. The sending unit is used to send the warning information to nearby vehicles, wherein the nearby vehicles are vehicles whose distance from the target area is less than or equal to a preset distance range.

10. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the method as described in any one of claims 1 to 5.

Images