Risk early warning method of vehicle, vehicle and electronic device

Abstract

The embodiment of the specification provides a risk early warning method of a vehicle, relates to the technical field of intelligent sensing, and the method comprises the following steps: acquiring at least one type of sensing information of a target vehicle in front of the vehicle; then judging a fire event of the target vehicle in front of the vehicle according to various sensing information, and sending early warning information to the target vehicle in front of the vehicle through a special communication protocol. The method realizes active sensing of an external vehicle fire event, extends fire monitoring capability from a self vehicle to other vehicles, and improves the timeliness of discovering a fire hazard; then the early warning information is sent to the target vehicle in front of the vehicle through the designated communication protocol, a directional early warning channel from a rear vehicle to a front vehicle is established, and the accuracy and reachability of early warning information transmission are improved; then the early warning operation of the front vehicle is triggered through the early warning information, the driver of the front vehicle is ensured to know the danger in the early stage of the fire, the timeliness and initiative of emergency response are improved, and therefore the safety of vehicle driving is improved.

Description

Technical Field

[0001] This specification relates to the field of intelligent sensing technology, specifically to vehicle risk warning technology in the field of intelligent sensing technology, and more specifically to a vehicle risk warning method, vehicle, and electronic equipment. Background Technology

[0002] Vehicle fires are frequent in modern road traffic, especially in high-speed driving and large truck cargo transportation scenarios. Drivers often cannot detect vehicle fires in time, which often leads to casualties and significant property damage.

[0003] Generally, fire warnings for vehicles are based on the temperature sensors installed in the vehicle. However, fires in the early stages of an accident or in blind spots of the vehicle's detection system are easily overlooked, affecting the safety of driving. Summary of the Invention

[0004] This specification provides a vehicle risk warning method, a vehicle, and electronic equipment to improve vehicle driving safety.

[0005] To achieve the above technical objectives, the embodiments of this specification provide the following technical solutions: Firstly, one embodiment of this specification provides a vehicle risk warning method, comprising: Acquire at least one type of perception information of a target vehicle in front of the vehicle, wherein the vehicle and the target vehicle in front of the vehicle are interconnected via a specified communication protocol, the specified communication protocol is configured with a message type identifier, and the message type identifier contains a standard field indicating warning information; Based on the various sensed information, determine whether a target vehicle in front of the vehicle has caught fire, so as to obtain event information; If the event information indicates that a fire has occurred in a target vehicle ahead of the vehicle, a warning message is sent to the target vehicle ahead of the vehicle through the specified communication protocol. The warning message is used to trigger the target vehicle ahead of the vehicle to perform a warning operation in response to the fire event.

[0006] Optionally, in one possible implementation, acquiring at least one type of perception information about a target vehicle ahead of the vehicle includes: Visual acquisition is performed based on the visual sensors configured in the vehicle to obtain image information about the target vehicle in front of the vehicle; Thermal imaging is performed based on the infrared sensors configured in the vehicle to obtain thermal information about the target vehicle in front of the vehicle. Particle density detection is performed based on the particulate matter sensor configured in the vehicle to obtain concentration information about the target vehicle in front of the vehicle; The image information, thermal information, and concentration information are aligned in time dimension by using timestamps to obtain perception information about the target vehicle in front of the vehicle.

[0007] In this implementation, by simultaneously configuring a visual sensor, an infrared sensor, and a particulate matter sensor, image information, thermal information, and concentration information of the target vehicle ahead are acquired respectively. This enables comprehensive acquisition of three complementary types of information: flame texture, temperature anomaly, and smoke particles, thereby improving the comprehensiveness of the perception dimensions. Furthermore, the image information, thermal information, and concentration information are time-stamped to accurately synchronize the data from the three heterogeneous sensors in the time dimension, providing a spatiotemporally consistent data foundation for subsequent multimodal fusion, thereby improving the availability of perception information and the accuracy of fusion.

[0008] Optionally, in one possible implementation, determining whether a target vehicle in front of the vehicle has caught fire based on various perception information to obtain event information includes: Feature extraction is performed based on the modal data distribution corresponding to the perceived information to obtain the perceived features, which include at least one of the following: image features of the flame, data distribution features in the thermal map, and data distribution features in the particle density map. The perceived features are respectively input into the corresponding attention branches for feature enhancement to obtain joint features, and the attention branches intersect with each other; Based on the combined features, a fire event is determined for the target vehicle in front of the vehicle to obtain the event information.

[0009] In this implementation, channel alignment is performed on data from different modalities in the perceived information to enable the processing of visual images, heatmaps, and particle densities. Figure 3 Heterogeneous data are standardized into a model input format to improve the operability of data fusion. Then, feature extraction is performed according to the distribution of each modality of data to obtain flame image features, heat map distribution features, and particle density distribution features, deeply mining the semantic information of each modality and improving the richness of feature expression. The extracted perceptual features are then input into the attention branches that intersect each other for feature enhancement, enabling cross-modal interaction and collaborative enhancement of features under the attention mechanism. This allows the model to automatically learn the optimal fusion weights among multiple modalities, improving the discriminative ability and anti-interference ability of the fused features. Finally, fire events are judged based on joint features, realizing fire identification under deep fusion of multimodal information, overcoming the perceptual limitations of a single modality in a specific scenario, and thus significantly improving the accuracy and robustness of fire event judgment.

[0010] Optionally, in one possible implementation, the method further includes: The joint features are input into the convolutional layer associated with the attention branch to perform a full connection on the joint features to obtain global features; The global features are mapped onto image features to obtain a confidence map indicating the location of the fire; The confidence map is used to mark the location of the fire on the target vehicle in front of the vehicle.

[0011] In this implementation, highly abstract global features are obtained by fully connecting the convolutional layers associated with the attention branch of the joint feature input, achieving overall semantic condensation of multimodal fusion information and improving the information density of feature representation. The global features are then mapped to the image feature space to generate a confidence map with the same resolution as the original image, completing the fine transformation from global judgment to pixel-level localization and improving the spatial accuracy of fire location. Then, the fire location is marked by the confidence map, providing accurate location information for subsequent targeted early warning, thereby improving the relevance and practicality of the early warning information.

[0012] Optionally, in one possible implementation, marking the location of the fire of the target vehicle in front of the vehicle using the confidence map includes: Obtain the video stream associated with the confidence map; Based on the video stream, determine a plurality of adjacent reference video frames of the confidence map; If the confidence map corresponding to the reference video frame meets the preset conditions, then the fire location of the target vehicle in front of the vehicle indicated by the confidence map is marked.

[0013] In this implementation, the video stream associated with the confidence map is obtained, and multiple adjacent reference video frames are determined based on the video stream. A judgment window based on the time sequence dimension is constructed, extending single-point judgment to time sequence analysis and improving the reliability of the judgment mechanism. Then, the confidence map corresponding to the reference video frame is judged under preset conditions. The marking is confirmed only when multiple consecutive frames meet the fire conditions. The temporal consistency is used to filter out occasional noise interference in a single frame, thereby improving the stability and accuracy of the fire location marking and effectively avoiding early warning errors caused by instantaneous misjudgment.

[0014] Optionally, in one possible implementation, if the event information indicates that a target vehicle ahead of the vehicle has caught fire, then sending a warning message to the target vehicle ahead of the vehicle via the specified communication protocol includes: If the event information indicates that a target vehicle in front of the vehicle has caught fire, then the vehicle identifier corresponding to the target vehicle in front of the vehicle is determined. The intensity level of the fire event is determined based on the event information; Configure location parameters based on fire location markers on confidence maps; Configure action codes to indicate warning actions based on the location parameters and intensity levels; The warning information is obtained by configuring a message type identifier using the vehicle identification, intensity level, location parameters, and motion code. Warning information is sent to the target vehicle ahead of the vehicle through the specified communication protocol.

[0015] In this implementation, the vehicle identification of the target vehicle ahead is determined to achieve directional addressing of the early warning information, ensuring that the early warning message is accurately delivered to the target vehicle, thus improving the targeting and privacy protection capabilities of the communication. The intensity level of the fire event is determined based on the event information, and location parameters are configured based on the fire location marked by the confidence map. This structures the fire severity and location information obtained from the front end, upgrading the early warning information from a simple alarm to a refined message containing multi-dimensional semantics, increasing the information content of the early warning information. Then, action codes are configured based on the location parameters and intensity level to achieve pre-planning of response strategies based on the specific fire situation, improving the executability and guidance value of the early warning information. Finally, message type identifiers are configured through vehicle identification, intensity level, location parameters, and action codes to form a standardized dedicated fire early warning message structure. This provides a standard data format for the application of specified communication protocols in fire emergency scenarios, improving the system's interoperability and standardization level.

[0016] Optionally, in one possible implementation, configuring the action coding by combining the position parameters and the intensity level includes: Configure the first action by combining the location parameters and the intensity level; Obtain environmental information corresponding to the target vehicle in front of the vehicle; The first action is adjusted based on the environmental information to obtain the second action; The action code is configured according to the second action.

[0017] In this implementation, by combining location parameters and intensity levels to configure the first action, a preliminary response strategy based on the fire situation is generated, directly linking the action suggestions to the fire location and severity, thus improving the targeting of the response suggestions. Furthermore, environmental information corresponding to the target vehicle ahead is acquired, and external variables such as traffic flow, weather conditions, and road type are introduced to enable environmentally adaptive action adjustments, improving the dynamic adaptability of the response strategy. Then, based on the environmental information, the first action is adjusted to obtain the second action, allowing the final action suggestion to comprehensively consider both the fire situation and the current scenario, thereby improving the safety and feasibility of the response suggestion.

[0018] Secondly, one embodiment of this specification provides a vehicle risk warning method, including: The system acquires warning information sent by a vehicle, wherein the vehicle is interconnected with a target vehicle in front of the vehicle via a specified communication protocol, and the warning information is used to indicate that a fire has occurred in the target vehicle in front of the vehicle. The fire event is determined based on at least one type of perception information of the target vehicle in front of the vehicle. The perception information is collected by a target sensor configured in the vehicle. The specified communication protocol is configured with a message type identifier, and the message type identifier contains a field indicating the warning information. In response to the receipt of the warning information, the processing action indicated by the warning information is triggered.

[0019] In this implementation, by acquiring the warning information sent by the vehicle, the target vehicle in front of the vehicle is established to receive cross-vehicle fire warnings, enabling the vehicle in front to receive active alarms from the vehicle behind, breaking the limitation that traditional vehicles cannot receive external fire signals, and improving the vehicle's ability to perceive external hazards; and responding to the receiving warning information triggers processing actions, realizing direct linkage between the warning information and the vehicle's actuators, enabling the vehicle in front to automatically complete various warning operations, thereby improving the automation and immediacy of emergency response, and ensuring that the driver is aware of the danger and takes action in the shortest possible time.

[0020] Thirdly, one embodiment of this specification provides a vehicle risk warning device, comprising: An acquisition unit is used to acquire at least one type of perception information of a target vehicle in front of the vehicle. The vehicle and the target vehicle in front of the vehicle are interconnected through a specified communication protocol. The specified communication protocol is configured with a message type identifier, and the message type identifier contains a standard field indicating warning information. The early warning unit is used to determine whether a target vehicle in front of the vehicle has caught fire based on various sensing information, so as to obtain event information; The warning unit is further configured to send warning information to the target vehicle in front of the vehicle via the specified communication protocol if the event information indicates that a fire has occurred in the target vehicle in front of the vehicle. The warning information is used to trigger the target vehicle in front of the vehicle to perform a warning operation in response to the fire event.

[0021] Fourthly, one embodiment of this specification also provides a vehicle risk warning system, including sensors and a vehicle controller, wherein the sensors are used to collect vehicle-related data; and the vehicle controller is used to execute the vehicle risk warning method as described in the first aspect or any possible implementation thereof.

[0022] Fifthly, one embodiment of this specification also provides a vehicle, the vehicle comprising: a memory for storing executable program code; and a processor for calling and running the executable program code from the memory, causing the vehicle to execute the vehicle risk warning method in the first aspect or any possible implementation of the first aspect.

[0023] Sixthly, one embodiment of this specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the vehicle risk warning method described above.

[0024] Seventhly, embodiments of this specification provide a computer program product or computer program, the computer program product including a computer program that can be stored in a computer-readable storage medium or in the cloud; the processor of the computer device reads the computer program, and when the processor executes the computer program, it implements the steps of the above-described vehicle risk warning method.

[0025] As can be seen from the above technical solution, the vehicle risk warning method provided in this specification obtains at least one type of perception information of a target vehicle in front of the vehicle. The vehicle and the target vehicle in front of the vehicle are interconnected through a specified communication protocol, and the specified communication protocol is configured with a message type identifier. The message type identifier contains a standard field indicating warning information. Then, based on various perception information, it is determined whether a fire event has occurred in the target vehicle in front of the vehicle to obtain event information. If the event information indicates that a fire event has occurred in the target vehicle in front of the vehicle, a warning information is sent to the target vehicle in front of the vehicle through the specified communication protocol. The warning information is used to trigger the target vehicle in front of the vehicle to perform a warning operation in response to the fire event. By acquiring different types of sensory information about target vehicles ahead, the system proactively detects external vehicle fires, extending fire monitoring capabilities from the user's own vehicle to other vehicles, thus improving the timeliness of fire hazard detection. Furthermore, by judging fire events in target vehicles ahead based on sensory information, the system achieves automated fire event recognition, transforming raw sensory data into actionable event information, thereby enhancing the intelligence level of information processing. Then, by sending warning information to target vehicles ahead via a designated communication protocol, a directional warning channel is established from the following vehicle to the preceding vehicle, improving the accuracy and reach of warning information transmission. Finally, the warning information triggers warning actions in the target vehicles ahead, ensuring that the driver of the preceding vehicle is aware of the danger in the early stages of a fire, improving the timeliness and proactivity of emergency response, and ultimately enhancing vehicle driving safety. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this specification. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0027] Figure 1 This is a schematic diagram illustrating the application environment of a vehicle risk warning method provided as one embodiment of this specification.

[0028] Figure 2 This is a flowchart illustrating a vehicle risk warning method according to one embodiment of this specification.

[0029] Figure 3 This is a schematic diagram of a vehicle risk warning method provided as one embodiment of this specification.

[0030] Figure 4 A schematic diagram of the functional modules of a vehicle risk warning device provided in one embodiment of this specification.

[0031] Figure 5 This is a structural schematic diagram of a vehicle provided for one embodiment of this specification. Detailed Implementation

[0032] Unless otherwise defined, the technical or scientific terms used in the embodiments of this specification shall have the meaning understood by one of ordinary skill in the art to which this specification pertains. The terms "first," "second," and similar terms used in the embodiments of this specification do not indicate any order, quantity, or importance, but are merely used to avoid confusion of constituent elements.

[0033] Unless the context otherwise requires, throughout this specification, "a plurality of" means "at least two," and "including" is interpreted as open-ended or encompassing, that is, "including, but not limited to." In the description of this specification, terms such as "one embodiment," "some embodiments," "exemplary embodiment," "example," "specific example," or "some examples" are intended to indicate that a particular feature, structure, material, or characteristic associated with that embodiment or example is included in at least one embodiment or example of this specification. The illustrative representations of the above terms do not necessarily refer to the same embodiment or example.

[0034] First, let's explain the possible terms: Vehicle-to-Vehicle (V2V) communication refers to the exchange of information between vehicles via Dedicated Short Range Communication (DSRC) or C-V2X technology.

[0035] YOLOv8 (You Only Look Once version 8): A high-efficiency real-time object detection deep learning model suitable for vehicle vision recognition.

[0036] Transformer: A deep learning architecture based on self-attention mechanism, which is good at modeling long-distance dependencies and is widely used in natural language processing and vision tasks. In this solution, it is used for cross-modal feature association.

[0037] Vehicle-mounted telematics box (T-Box): responsible for communication, positioning and remote control.

[0038] Head-Up Display (HUD): Projects key information onto the windshield, reducing the need for drivers to look down.

[0039] European Telecommunications Standards Institute (ETSI): One of the main organizations that develop V2X communication standards.

[0040] Automotive network security and message format standard (IEEE 1609.2): Used for V2X communication security authentication and data encapsulation.

[0041] Protocol for synchronizing network device clocks (IEEE 1588 PTP): A protocol used to synchronize the clocks of network devices, especially in local area networks (LANs), with the goal of achieving sub-microsecond clock synchronization within the network.

[0042] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0043] Vehicle fires are frequent in modern road traffic, especially in high-speed driving and large truck cargo transportation scenarios. Drivers often cannot detect vehicle fires in time, which often leads to casualties and significant property damage.

[0044] Generally, fire warnings for vehicles are based on the temperature sensors installed in the vehicle. However, fires in the early stages or in blind spots of the vehicle's detection system are easily overlooked. In other words, when a vehicle catches fire (such as battery thermal runaway, cargo box fire, or tire overheating and smoke), there is a lack of onboard automatic detection and alarm mechanisms. Drivers often only notice the fire after it has spread, missing the best window for handling the situation and affecting the safety of driving the vehicle.

[0045] To address the aforementioned problems, this specification provides a vehicle risk warning system, and the vehicle risk warning method provided in this specification is applied to the vehicle risk warning system. This warning system deploys a multimodal fire detection module in the following vehicle and broadcasts a dedicated fire warning message to the target preceding vehicle based on a specified communication protocol. Upon receiving the message, the preceding vehicle automatically triggers different priority alert processes, thus achieving an accurate risk warning process.

[0046] Specifically, the vehicle's risk warning system may include a system composed of... Figure 1 The system comprises a client 110, a server 120, and a vehicle 130, forming an operating environment. The client 110 communicates with the server 120 via a network. The client 110 is wirelessly connected to the vehicle 130. The vehicle 130 communicates with the server 120 via a network connection. The client 110 can be an electronic device with network access capabilities. Specifically, for example, the client 110 can be a desktop computer, tablet, laptop, smartphone, digital assistant, smart wearable device, shopping guide terminal, television, smart speaker, microphone, etc. Smart wearable devices include, but are not limited to, smart bracelets, smartwatches, smart glasses, smart helmets, smart necklaces, etc. Alternatively, the client 110 can also be software that can run on the electronic device. The server 120 can be an electronic device with certain computing power. It can have a network communication module, processor, and memory, etc. Of course, the server 120 can also refer to software running on the electronic device. The server 120 can also be a distributed server, which can be a system with multiple processors, memory, network communication modules, etc., operating collaboratively. Alternatively, the server 120 can also be a cluster formed by several servers 120. Alternatively, with the development of science and technology, server 120 could also be a new technological means capable of realizing the corresponding functions of the implementation method described in the manual. For example, it could be a new form of "server" based on quantum computing.

[0047] Specifically, when the aforementioned early warning system performs risk identification inside the vehicle, it acquires at least one type of perception information about the target vehicle in front of the vehicle. The vehicle and the target vehicle in front of the vehicle are interconnected through a specified communication protocol, and the specified communication protocol is configured with a message type identifier. The message type identifier contains a standard field indicating early warning information. Then, it determines whether a fire event has occurred in the target vehicle in front of the vehicle based on various perception information to obtain event information. If the event information indicates that a fire event has occurred in the target vehicle in front of the vehicle, it sends early warning information to the target vehicle in front of the vehicle through the specified communication protocol. The early warning information is used to trigger the target vehicle in front of the vehicle to perform an early warning operation in response to the fire event. By acquiring different types of sensory information about target vehicles ahead, the system proactively detects external vehicle fires, extending fire monitoring capabilities from the user's own vehicle to other vehicles, thus improving the timeliness of fire hazard detection. Furthermore, by judging fire events in target vehicles ahead based on sensory information, the system achieves automated fire event recognition, transforming raw sensory data into actionable event information, thereby enhancing the intelligence level of information processing. Then, by sending warning information to target vehicles ahead via a designated communication protocol, a directional warning channel is established from the following vehicle to the preceding vehicle, improving the accuracy and reach of warning information transmission. Finally, the warning information triggers warning actions in the target vehicles ahead, ensuring that the driver of the preceding vehicle is aware of the danger in the early stages of a fire, improving the timeliness and proactivity of emergency response, and ultimately enhancing vehicle driving safety.

[0048] Based on the above concept, this specification provides a vehicle risk warning method. The vehicle risk warning method provided in this specification will be described exemplarily below with reference to the accompanying drawings.

[0049] To be applied Figure 1 Taking a vehicle as an example, this specification provides illustrative examples of some embodiments of the risk warning method for that vehicle. Figure 2 As shown, Figure 2 A flowchart illustrating a vehicle risk warning method according to one embodiment of this specification; the vehicle risk warning method includes: 201. Acquire at least one type of perception information of a target vehicle in front of the vehicle. The vehicle and the target vehicle in front of the vehicle are interconnected through a specified communication protocol. The specified communication protocol is configured with a message type identifier, which contains a standard field indicating warning information.

[0050] In this embodiment, the vehicle refers to the vehicle responsible for sensing and sending early warnings, which can be a vehicle traveling behind or to the side of the target vehicle. The vehicle is equipped with a multimodal sensor array and an onboard communication module, responsible for environmental perception of the target vehicle ahead, fire event assessment, and, upon confirmation of a fire event, proactively sending early warning information to the target vehicle ahead via a specified communication protocol. In this solution, the vehicle acts as both an "observer" and an "alarmer," actively identifying potential fire hazards in external vehicles and disseminating the information accordingly, thus achieving the information source function for cross-vehicle fire early warning.

[0051] Correspondingly, the target vehicle in front of the vehicle refers to the vehicle responsible for receiving and responding to early warnings, which can be a target vehicle traveling in front of or around the vehicle. The target vehicle in front of the vehicle is equipped with an onboard communication module and an early warning response unit, capable of receiving dedicated early warning information from other vehicles and automatically triggering in-vehicle audible and visual alarms, HUD prompts, and auxiliary braking based on the fire intensity, fire location, and suggested actions in the early warning information. By receiving fire alarms from external vehicles and responding promptly in the early stages of a fire, the target vehicle in front of the vehicle effectively improves the timeliness and safety of fire emergency response.

[0052] It is understandable that the positional relationship between the vehicle and the target vehicle in front of the vehicle can be front-to-back, left-to-right, or other possible sides. The specific positional relationship depends on the actual scenario. In this embodiment, the vehicle is the rear vehicle and the target vehicle in front of the vehicle is the front vehicle.

[0053] Specifically, the vehicle perception process can involve various types of perceptual information, which can be configured based on different data modalities. Different models require sensors of different dimensions. Among these, visual sensors refer to visible light imaging devices installed on the vehicle, such as high-resolution cameras mounted on the unobstructed inner side of the windshield, to acquire RGB image information of the target vehicle in front. Visual sensors provide basic data for visual recognition of fire events by capturing the texture and color features of flames and the morphological features of smoke. In fire perception, visual sensors primarily perform "morphological recognition," intuitively reflecting the outline and color of flames and the diffusion pattern of smoke. However, their perception capabilities are limited at night, in rainy or foggy weather, or when visibility is obstructed, requiring collaboration with other sensors.

[0054] Infrared sensors are thermal imaging devices installed on vehicles. They can be narrow-band infrared thermal imagers, mounted rigidly on the inside of the windshield and sharing a base with the visual sensor. They are used to collect thermal image information about target vehicles in front of the vehicle. By detecting the infrared radiation emitted by the target object, the infrared sensor generates a thermal map reflecting the temperature distribution, accurately identifying early-stage fire conditions such as battery thermal runaway, tire overheating, and abnormal temperature rise in the engine compartment, even without open flames. In fire detection, infrared sensors primarily perform the "temperature detection" function. They can penetrate smoke and are unaffected by lighting conditions, effectively compensating for the limitations of visual sensors at night and in situations where visibility is obstructed.

[0055] A particulate matter sensor is a smoke particle detection device installed on a vehicle. It utilizes the principle of laser scattering, is mounted inside the front bumper, and passively samples the airflow to collect information about the smoke concentration in the surrounding environment of the target vehicle. By detecting the concentration of suspended particulate matter in the air with a diameter of 0.3–10 μm, the particulate matter sensor can capture particles produced by combustion and has extremely high sensitivity in the early stages of smoldering fires. In fire detection, the particulate matter sensor primarily performs the function of "smoke detection," providing early warning of fire risks by detecting abnormal increases in smoke particle concentration before visual sensors detect open flames or infrared sensors detect significant temperature rises.

[0056] It is understandable that the aforementioned sensors, with their different dimensions, characterize fire events from different perspectives, and one or more of them can be used in combination. The following embodiment illustrates the process of multi-source data fusion.

[0057] By configuring sensors across different dimensions as described above, perception information can be obtained. Perception information refers to multimodal environmental data about the target vehicle ahead, acquired and processed by the vehicle through the configured target sensors. This perception information includes image information acquired by visual sensors, thermal information acquired by infrared sensors, and smoke concentration information acquired by particulate matter sensors. These three types of data are aligned using timestamps to form a spatiotemporally consistent fusion perception result. Perception information forms the original data foundation for fire detection in this technical solution, encompassing the visual characteristics of flames, the spatial distribution of temperature, and changes in smoke concentration, providing comprehensive and multidimensional input for subsequent deep learning models.

[0058] Specifically, the process of acquiring multidimensional perception information can involve visual acquisition based on the vehicle's configured visual sensors to obtain image information about the target vehicle in front of the vehicle; thermal imaging based on the vehicle's configured infrared sensors to obtain thermal information about the target vehicle in front of the vehicle; particle density detection based on the vehicle's configured particulate sensors to obtain concentration information about the target vehicle in front of the vehicle; and finally, time-dimensional alignment of the image information, thermal information, and concentration information using timestamps to obtain perception information about the target vehicle in front of the vehicle.

[0059] Furthermore, communication between a vehicle and a target vehicle ahead is based on a specified communication protocol. This protocol is a standardized communication mechanism specifically designed for transmitting fire warning information between vehicles. It is implemented by extending the IEEE 1609.2 vehicle-to-everything (V2X) communication standard framework, and can therefore also be called a dedicated communication protocol. This protocol defines dedicated message type identifiers (such as V2V_FIRE_ALERT), and its payload structure includes fire emergency-specific fields such as target vehicle identification, fire intensity level, fire location enumeration value, and suggested action code. It also sets the highest communication priority to ensure message delivery within 500ms. This dedicated communication protocol differs from general V2X location exchange or emergency braking broadcast protocols, providing a standardized message format and transmission mechanism for the specific scenario of "a following vehicle reporting a fire to the vehicle ahead."

[0060] In one possible scenario, the aforementioned perception process can be configured within a multi-source sensor fusion module integrated at the rear of the vehicle. Specifically, a high-resolution visible light camera (resolution ≥1920×1080, frame rate ≥30fps, used for acquiring RGB images) and a narrow-band infrared thermal imager (band coverage 8–10μm, resolution ≥320×240, NETD≤50mK, used for generating thermal maps) are fixedly installed in the unobstructed area inside the windshield. Both cameras achieve parallel optical axis calibration through a rigid shared base structure, ensuring image spatial alignment error ≤0.5 pixels. A miniature smoke particle sensor (using laser scattering principle, detection particle size range 0.3–10μm, sampling frequency ≥10Hz, volume ≤20cm³, protection rating IP67, used for collecting smoke concentration) is embedded inside the front bumper, with its air intake facing the vehicle's direction of travel, utilizing airflow-guided passive sampling.

[0061] Furthermore, all sensor data is uniformly accessed to the domain controller via the vehicle Ethernet, and the timestamp synchronization protocol (IEEE 1588 PTP) is used to achieve microsecond-level time alignment of visual, infrared, and particulate matter data, and the three types of data are mapped to the same vehicle coordinate system.

[0062] As can be seen, in this embodiment, by simultaneously configuring a visual sensor, an infrared sensor, and a particulate matter sensor, image information, thermal information, and concentration information of the target vehicle in front of the vehicle are acquired respectively, achieving comprehensive collection of three complementary types of information: flame texture, temperature anomaly, and smoke particles, thereby improving the comprehensiveness of the perception dimensions. Furthermore, the image information, thermal information, and concentration information are time-stamped to accurately synchronize the data from the three heterogeneous sensors in the time dimension, providing a spatiotemporally consistent data foundation for subsequent multimodal fusion, thereby improving the usability of the perception information and the fusion accuracy.

[0063] 202. Determine whether a target vehicle in front of the vehicle has caught fire based on various sensory information in order to obtain event information.

[0064] In this embodiment, event information refers to the structured result of the fire status of a target vehicle ahead, output after judgment is made based on perceived information using a deep learning model. Event information includes at least a binary classification result indicating whether a fire event has occurred, and may further include fire intensity levels (e.g., smoke, small fire, large fire), fire location markers (e.g., battery compartment, cargo box, tires), and a pixel-level confidence map of the fire location. Event information is a crucial intermediate result in transforming raw perceived data into executable instructions, representing an intelligent leap from environmental perception to event recognition, and providing core data support for the generation of subsequent early warning information.

[0065] Specifically, the determination of a fire event can be performed by constructing a deep learning fire determination model. This involves first acquiring channel-aligned perception information and then extracting features based on the modal data distribution corresponding to the perception information to obtain perception features. These perception features include at least one of the following: image features of the flame, data distribution features in the heat map, and data distribution features in the particle density map. The perception features are then input into the corresponding attention branches for feature enhancement to obtain joint features, with the attention branches intersecting with each other. Finally, based on the joint features, a fire event is determined for the target vehicle in front of the vehicle to obtain event information.

[0066] The process of channel alignment involves designing a three-channel multimodal input structure: the images (RGB) captured by the visible light camera, the thermal map (TIR) ​​output by the infrared thermal imager, and the particle density map (SmokeMask) generated by the smoke particle sensor are uniformly normalized into a 256×256×3 tensor, which is used as the unified input of the network to achieve structural alignment of the physical sensor data.

[0067] Furthermore, the feature extraction process involves constructing a YOLOv8-Transformer hybrid architecture: using YOLOv8 as the backbone extractor, it extracts flame texture and color features from RGB images, morphological and gradient features of high-temperature regions from heatmaps, and smoke diffusion patterns and aggregation density features from particle density maps, performing local spatial feature encoding and outputting three sets of feature maps. These three sets of feature maps are then concatenated along the channel dimension and input into the Transformer module. This module contains three sets of cross-attention heads, which calculate cross-modal attention weights between RGB and TIR, RGB and Smoke, and TIR and Smoke, respectively, achieving spatial semantic alignment and feature enhancement of flame outline (RGB), high-temperature distribution (TIR), and smoke diffusion (Smoke). The three sets of cross-attention outputs are then concatenated and channel compressed to generate a unified joint feature vector of "flame outline + high temperature + smoke," and the joint feature vector is classified into three fire intensity categories: "smoke," "small fire," and "large fire."

[0068] As can be seen, this embodiment aligns the channels of data from different modalities in the perceived information, enabling visual images, heatmaps, and particle densities to be integrated. Figure 3 Heterogeneous data are standardized into a model input format to improve the operability of data fusion. Then, feature extraction is performed according to the distribution of each modality of data to obtain flame image features, heat map distribution features, and particle density distribution features, deeply mining the semantic information of each modality and improving the richness of feature expression. The extracted perceptual features are then input into the attention branches that intersect each other for feature enhancement, enabling cross-modal interaction and collaborative enhancement of features under the attention mechanism. This allows the model to automatically learn the optimal fusion weights among multiple modalities, improving the discriminative ability and anti-interference ability of the fused features. Finally, fire events are judged based on joint features, realizing fire identification under deep fusion of multimodal information, overcoming the perceptual limitations of a single modality in a specific scenario, and thus significantly improving the accuracy and robustness of fire event judgment.

[0069] Furthermore, to facilitate fire risk assessment, the location can be marked by inputting the joint features into a convolutional layer associated with the attention branch to perform a full connection on the joint features to obtain global features; mapping the global features onto image features to obtain a confidence map indicating the fire location; and marking the fire location of the target vehicle in front of the vehicle using the confidence map.

[0070] This process is achieved by introducing a position-aware confidence head, which directly connects an independent 1×1 convolutional layer (equivalent to a fully connected head) to the end of the Transformer output. This directly maps the extracted global features into a "fire location confidence map" with the same resolution as the input image, which is used to accurately locate the fire area (battery compartment, vehicle body, tires, etc.) and support subsequent visualization and linkage response.

[0071] As can be seen, this embodiment obtains highly abstract global features by fully connecting the convolutional layers associated with the attention branch of the joint feature input, achieving overall semantic condensation of multimodal fusion information and improving the information density of feature representation; and maps the global features to the image feature space to generate a confidence map with the same resolution as the original image, completing the fine transformation from global judgment to pixel-level localization, improving the spatial accuracy of fire location; then, the fire location is marked by the confidence map, providing accurate location information for subsequent targeted early warning, thereby improving the relevance and practicality of the early warning information.

[0072] In one possible scenario, in order to eliminate noise interference in the confidence level, a continuous frame dynamic determination logic can also be performed. That is, firstly, the video stream associated with the confidence level map is obtained; then, multiple reference video frames adjacent to the confidence level map are determined based on the video stream; if the confidence level map corresponding to the reference video frame meets the preset conditions, the fire position of the target vehicle in front of the vehicle indicated by the confidence level is marked.

[0073] The preset condition is that, within multiple consecutive reference video frames, the average confidence score of the fire location confidence map generated in each frame exceeds a set threshold. This condition utilizes the temporal continuity of the fire to filter out occasional noise in a single frame, ensuring that the marker is confirmed only when multiple frames continuously indicate a fire.

[0074] For example, in video stream processing, a sliding window can be set to cache the output of the "fire location confidence map" for the most recent 3 frames (multiple consecutive reference video frames). Only when the average value of the confidence map in the 3 consecutive frames is greater than 90% (a set threshold) will a valid fire event alarm be triggered; otherwise, it will be regarded as interference or noise.

[0075] As can be seen, this embodiment obtains the video stream associated with the confidence map and determines multiple adjacent reference video frames based on the video stream to construct a judgment window based on the time dimension, extending single-point judgment to time-series analysis and improving the reliability of the judgment mechanism. Then, the confidence map corresponding to the reference video frame is judged under preset conditions, and the mark is confirmed only when multiple consecutive frames meet the fire conditions. The temporal consistency is used to filter out the occasional noise interference of a single frame, thereby improving the stability and accuracy of the fire location marking and effectively avoiding warning errors caused by instantaneous misjudgment.

[0076] 203. If the event information indicates that a target vehicle in front of the vehicle has caught fire, a warning message is sent to the target vehicle in front of the vehicle through a specified communication protocol. The warning message is used to trigger the target vehicle in front of the vehicle to perform a warning operation in response to the fire event.

[0077] In this embodiment, the warning information refers to a standardized emergency message sent by the vehicle to the target vehicle ahead via a dedicated communication protocol after the vehicle determines that a fire has occurred in the target vehicle ahead. The warning information is encapsulated in the form of a dedicated message frame, and its payload includes at least the target vehicle identifier, fire intensity level, fire location parameters, and suggested action codes. Furthermore, the action suggestions can be dynamically adjusted based on the environmental information of the target vehicle ahead. The warning information carries a complete command stream from the sensing end to the response end, enabling the target vehicle ahead to automatically trigger audible and visual alarms, HUD prompts, and assisted braking operations upon receiving the message.

[0078] Specifically, the configuration process for early warning information can begin by first identifying the vehicle identifier corresponding to the target vehicle ahead; then determining the intensity level of the fire event based on the event information; configuring location parameters based on the fire location marked by the confidence map; configuring action codes to indicate early warning actions by combining the location parameters and intensity level; then configuring message type identifiers through vehicle identifier, intensity level, location parameters, and action codes to obtain early warning information; and finally sending the early warning information to the target vehicle ahead through a dedicated communication protocol.

[0079] Action coding represents standardized emergency response recommendations for a fire involving a vehicle ahead. Specifically, it is a set of predefined values ​​or instructions used to inform the driver of the vehicle ahead or the autonomous driving system what actions should be taken to maximize safety. For example, action coding can correspond to different response plans such as "immediately pull over," "slow down and turn on hazard lights," "maintain a safe distance and call for help," or "emergency lane change to the emergency lane."

[0080] For example, a new V2V_FIRE_ALERT message type identifier (msgType = 0x0F) is added to the SecurityMessage structure of IEEE 1609.2, and its Payload structure is defined to conform to the message coding specification of ETSI TS 103 097 to ensure compatibility with existing infrastructure. The “V2V_FIRE_ALERT” message type includes: a 64-bit front vehicle ID (a unique authentication identifier associated with the vehicle license plate based on RGB image recognition), a 2-bit fire intensity level (0=low (smoke), 1=medium (small fire), 2=high (large fire)), a 2-bit fire location enumeration value (0=battery compartment, 1=car compartment, 2=tire), and an 8-bit suggested action code (0=immediately pull over, 1=slow down and turn on hazard lights, 2=maintain distance and call for help, 3=emergency lane change to emergency lane).

[0081] As can be seen, this embodiment achieves directional addressing of early warning information by determining the vehicle identification of the target vehicle in front of the vehicle, ensuring that the early warning message is accurately delivered to the target vehicle, and improving the targeting and privacy protection capabilities of the communication. Furthermore, it determines the intensity level of the fire event based on event information, and configures location parameters based on the fire location marked by the confidence map, structuring the fire severity and location information obtained by the front end perception. This upgrades the early warning information from a simple alarm to a refined message containing multi-dimensional semantics, increasing the information content of the early warning information. Then, it combines location parameters and intensity level to configure action codes, realizing the pre-planning of response strategies based on the specific fire situation, improving the executability and guidance value of the early warning information. Finally, by configuring message type identifiers through vehicle identification, intensity level, location parameters, and action codes, a standardized dedicated fire early warning message structure is formed, providing a standard data format for the application of dedicated communication protocols in fire emergency scenarios, improving the system's interoperability and standardization level.

[0082] Furthermore, to improve the accuracy of risk assessment, environmental factors can also be considered. Specifically, the first action is configured by combining location parameters and intensity level; environmental information corresponding to the target vehicle in front of the vehicle is obtained; the first action is then adjusted based on the environmental information to obtain the second action; and action coding is configured according to the second action.

[0083] For example, based on the fire location being "tire" and the intensity level being "small fire," the system configures the first action as "decelerate and turn on hazard lights." Simultaneously, the system obtains the current road type as a highway through the in-vehicle navigation system and the current traffic flow as dense through the V2X roadside unit. Considering the comprehensive environmental information, the first action is adjusted to "maintain speed and turn on hazard lights, find the nearest service area and leave." The adjusted action code is then written into the warning message.

[0084] As can be seen, this embodiment combines location parameters and intensity levels to configure the first action, thereby generating a preliminary response strategy based on the fire situation. This directly links the action suggestions to the location and severity of the fire, improving the targeting of the response suggestions. Furthermore, it acquires environmental information corresponding to the target vehicle in front of the vehicle, introducing external variables such as traffic flow, weather conditions, and road type, enabling the action adjustment to have environmental adaptability and improving the dynamic adaptability of the response strategy. Then, based on the environmental information, the first action is adjusted to obtain the second action, so that the final action suggestion can comprehensively consider both the fire situation and the current scenario, improving the safety and feasibility of the response suggestion.

[0085] In one possible scenario, the execution process for the target vehicle in front of the vehicle involves first acquiring the warning information sent by the vehicle. The vehicle and the target vehicle in front of the vehicle are interconnected through a dedicated communication protocol. The warning information is used to indicate that a fire event has occurred in the target vehicle in front of the vehicle. The fire event is determined based on the perception information of the target vehicle in front of the vehicle. The perception information is collected by the target sensors configured on the vehicle. The specified communication protocol is configured with a message type identifier, which contains a field indicating the warning information. Then, in response to the receipt of the warning information, the processing action indicated by the warning information is triggered.

[0086] Specifically, a dedicated receiving and responding unit can be configured at the front vehicle (the target vehicle in front of the vehicle). The vehicle-mounted T-Box module listens for the "V2V_FIRE_ALERT" message in the V2V network. Once it matches its own ID, it immediately activates a three-level response: ① A high-frequency alarm is sounded in the vehicle; ② The HUD can flash detailed prompts such as "Vehicle cargo box on fire, pull over immediately" and "Vehicle tires are smoking, slow down and turn on hazard lights" based on the location and intensity of the fire and handling suggestions; ③ ESP auxiliary braking is automatically activated.

[0087] It is understood that the above warning methods are only examples, and the specific warning configuration may vary depending on the vehicle's configuration.

[0088] In addition, a communication priority mechanism can be established for the communication process with the target vehicle in front of the vehicle. The message is set to the highest priority (Priority Level 0) in the V2X communication link, which can penetrate low-speed network congestion and ensure delivery within 500ms.

[0089] Furthermore, regarding the configuration of early warning information, the acquisition of sensing information, and the process of judging events, please refer to the relevant descriptions in steps 201-203 above, which will not be repeated here.

[0090] In one possible scenario, regarding the interaction process between the aforementioned vehicles, such as Figure 3 As shown, Figure 3This diagram illustrates a scenario of a vehicle risk warning method provided in one embodiment of this specification. The diagram shows a following vehicle (the vehicle behind) acting as a sensing end, using integrated visual sensors, infrared sensors, and particulate matter sensors to perform multimodal environmental perception of the preceding vehicle, collecting real-time information such as images, thermal distribution, and smoke concentration. This perceived data, after time alignment, is input into a deep learning model for fire event fusion and judgment, outputting event information including fire intensity, fire location, and confidence level. Once a valid fire event is determined, the following vehicle sends a warning message encapsulated with the target vehicle's identifier, fire intensity, fire location, and suggested actions to the preceding vehicle via a dedicated communication protocol. This dedicated protocol defines fire warning message types within the V2X framework and sets the highest communication priority to ensure rapid and reliable delivery of warning information.

[0091] In addition, the vehicle in front (the target vehicle in front of the vehicle) acts as the response end. After receiving the warning information, it automatically triggers the action response mechanism, including in-vehicle audible and visual alarms, HUD display of the specific fire location and handling suggestions, and activation of ESP auxiliary braking, etc., to achieve a complete closed loop from the rear vehicle's perception to the front vehicle's response, which significantly improves the timeliness and safety of fire emergency response.

[0092] In this embodiment, a "reverse warning for a fire ahead" message standard is defined using a dedicated V2V messaging protocol, filling a gap in V2X communication for fire emergency scenarios. Combined with visible light, infrared thermal imaging, and smoke particle detection, the false alarm rate is significantly reduced. A cross-attention mechanism dynamically weights the contributions of each modality, automatically identifying key feature combinations and classifying fire intensity and locating the fire location. Automatic triggering of audible and visual alarms and auxiliary braking by the vehicle system improves response efficiency. Since the message is sent only to the target vehicle ahead, broadcast storms are avoided, ensuring both communication efficiency and privacy security.

[0093] In summary, this embodiment acquires at least one type of perception information about a target vehicle ahead of the vehicle. The vehicle ahead and the target vehicle are interconnected via a specified communication protocol, which is configured with a message type identifier. The message type identifier contains a standard field indicating warning information. Then, based on various perception information, it is determined whether a fire event has occurred in the target vehicle ahead, so as to obtain event information. If the event information indicates that a fire event has occurred in the target vehicle ahead, a warning information is sent to the target vehicle ahead via the specified communication protocol. The warning information is used to trigger the target vehicle ahead to perform a warning operation in response to the fire event. By acquiring different types of sensory information about target vehicles ahead, the system proactively detects external vehicle fires, extending fire monitoring capabilities from the user's own vehicle to other vehicles, thus improving the timeliness of fire hazard detection. Furthermore, by judging fire events in target vehicles ahead based on sensory information, the system achieves automated fire event recognition, transforming raw sensory data into actionable event information, thereby enhancing the intelligence level of information processing. Then, by sending warning information to target vehicles ahead via a designated communication protocol, a directional warning channel is established from the following vehicle to the preceding vehicle, improving the accuracy and reach of warning information transmission. Finally, the warning information triggers warning actions in the target vehicles ahead, ensuring that the driver of the preceding vehicle is aware of the danger in the early stages of a fire, improving the timeliness and proactivity of emergency response, and ultimately enhancing vehicle driving safety.

[0094] It should be noted that the various embodiments described in this specification emphasize the parts that differ from other embodiments, and the embodiments can be explained by comparison with each other. Any combination of the various embodiments described in this specification based on general technical knowledge is covered within the scope of this specification.

[0095] In one exemplary embodiment of this specification, a vehicle risk warning device 400 is also provided, such as... Figure 4 As shown, Figure 4 This specification provides a functional module diagram of a vehicle risk warning device 400 according to one embodiment of the present invention. The warning device 400 includes: The acquisition unit 401 is used to acquire at least one type of perception information of a target vehicle in front of the vehicle. The vehicle and the target vehicle in front of the vehicle are interconnected through a specified communication protocol. The specified communication protocol is configured with a message type identifier, which contains a standard field indicating warning information. The early warning unit 402 is used to determine whether a target vehicle in front of the vehicle has caught fire based on various sensing information, so as to obtain event information; The warning unit 402 is also used to send a warning message to the target vehicle in front of the vehicle through the specified communication protocol if the event information indicates that a fire has occurred in the target vehicle in front of the vehicle. The warning message is used to trigger the target vehicle in front of the vehicle to perform a warning operation in response to the fire event.

[0096] Optionally, in one possible embodiment, the warning unit 402 is specifically used to perform visual acquisition based on the visual sensors configured in the vehicle to obtain image information about a target vehicle in front of the vehicle. The warning unit 402 is specifically used to perform thermal imaging based on the infrared sensors configured on the vehicle in order to obtain thermal information about the target vehicle in front of the vehicle. The warning unit 402 is specifically used to perform particle density detection based on the particulate matter sensor configured on the vehicle in order to obtain concentration information about the target vehicle in front of the vehicle. The warning unit 402 is specifically used to align the image information, thermal information and concentration information in the time dimension using timestamps to obtain perception information about the target vehicle in front of the vehicle.

[0097] Optionally, in one possible embodiment, the warning unit 402 is specifically used to extract features based on the modal data distribution corresponding to the sensing information to obtain sensing features, which include at least one of the image features of the flame, the data distribution features in the thermal map, and the data distribution features in the particle density map. The early warning unit 402 is specifically used to input the perceived features into the corresponding attention branches for feature enhancement to obtain joint features, and the attention branches intersect with each other; The warning unit 402 is specifically used to determine the fire event of the target vehicle in front of the vehicle based on the joint features, so as to obtain the event information.

[0098] Optionally, in one possible embodiment, the warning unit 402 is specifically used to input the joint feature into the convolutional layer associated with the attention branch to perform a full connection on the joint feature to obtain global features; The early warning unit 402 is specifically used to map the global features onto image features to obtain a confidence map indicating the location of the fire. The warning unit 402 is specifically used to mark the location of the fire of the target vehicle in front of the vehicle using the confidence map.

[0099] Optionally, in one possible embodiment, the warning unit 402 is specifically used to acquire the video stream associated with the confidence map; The early warning unit 402 is specifically used to determine multiple reference video frames adjacent to the confidence map based on the video stream; The warning unit 402 is specifically used to mark the location of the fire of the target vehicle in front of the vehicle indicated by the confidence map if the confidence map corresponding to the reference video frame meets the preset conditions.

[0100] Optionally, in one possible embodiment, the warning unit 402 is specifically used to determine the vehicle identifier corresponding to the target vehicle in front of the vehicle if the event information indicates that a target vehicle in front of the vehicle has caught fire. The early warning unit 402 is specifically used to determine the intensity level of the fire event based on the event information; The early warning unit 402 is specifically used to configure location parameters based on the fire location marked on the confidence map; The warning unit 402 is specifically used to configure an action code that indicates a warning action by combining the location parameter and the intensity level; The warning unit 402 is specifically used to obtain the warning information by configuring a message type identifier through the vehicle identification, intensity level, location parameters and action code; The warning unit 402 is specifically used to send warning information to the target vehicle in front of the vehicle through the specified communication protocol.

[0101] Optionally, in one possible embodiment, the warning unit 402 is specifically configured to combine the location parameter and the intensity level to configure a first action; The warning unit 402 is specifically used to obtain environmental information corresponding to the target vehicle in front of the vehicle. The early warning unit 402 is specifically used to adjust the first action based on the environmental information to obtain the second action; The warning unit 402 is specifically used to configure the action code according to the second action.

[0102] Specifically, the acquisition unit and the early warning unit in this embodiment can correspond to physical components. For example, the early warning unit can be a processing module such as a CPU, GPU, or FPGA. The specific physical component can be any component or combination of components with the above functions. The specific method depends on the actual scenario and is not limited here.

[0103] The aforementioned warning device acquires at least one type of perception information about a target vehicle ahead of the vehicle. The vehicle ahead and the target vehicle are interconnected via a specified communication protocol, which is configured with a message type identifier. The message type identifier contains a standard field indicating warning information. Then, based on various perception information, it determines whether a fire has occurred in the target vehicle ahead, thereby obtaining event information. If the event information indicates that a fire has occurred in the target vehicle ahead, a warning message is sent to the target vehicle ahead via the specified communication protocol. The warning message is used to trigger the target vehicle ahead to perform a warning operation in response to the fire event. By acquiring different types of sensory information about target vehicles ahead, the system proactively detects external vehicle fires, extending fire monitoring capabilities from the user's own vehicle to other vehicles, thus improving the timeliness of fire hazard detection. Furthermore, by judging fire events in target vehicles ahead based on sensory information, the system achieves automated fire event recognition, transforming raw sensory data into actionable event information, thereby enhancing the intelligence level of information processing. Then, by sending warning information to target vehicles ahead via a designated communication protocol, a directional warning channel is established from the following vehicle to the preceding vehicle, improving the accuracy and reach of warning information transmission. Finally, the warning information triggers warning actions in the target vehicles ahead, ensuring that the driver of the preceding vehicle is aware of the danger in the early stages of a fire, improving the timeliness and proactivity of emergency response, and ultimately enhancing vehicle driving safety.

[0104] Specific limitations regarding vehicle risk warning devices can be found in the above section on vehicle risk warning methods, and will not be repeated here. Each unit module in the aforementioned vehicle risk warning device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0105] In addition, in one exemplary embodiment of this specification, a vehicle is also provided, such as Figure 5 As shown, Figure 5 This is a structural schematic diagram of a vehicle provided for one embodiment of this specification.

[0106] For example, vehicle 500 and Figure 1 Vehicle 120 in the text refers to the same vehicle.

[0107] For example, such as Figure 5 As shown, the vehicle 500 includes a memory 510 and a processor 520, wherein the memory 510 stores executable program code 530, and the processor 520 is used to call and execute the executable program code 530 to perform a method for risk warning of a vehicle.

[0108] For example, the memory 510 can be used to store the relevant program of the vehicle risk warning method provided in the embodiments of this application; the processor 520 can call the relevant program of the vehicle risk warning method stored in the memory 510 to execute the vehicle risk warning method of the embodiments of this application; for example, obtaining at least one type of perception information of the target vehicle in front of the vehicle, the vehicle and the target vehicle in front of the vehicle are interconnected through a specified communication protocol, the specified communication protocol is configured with a message type identifier, the message type identifier contains a standard field indicating warning information; judging whether the target vehicle in front of the vehicle has caught fire based on various perception information to obtain event information; if the event information indicates that the target vehicle in front of the vehicle has caught fire, then sending warning information to the target vehicle in front of the vehicle through the specified communication protocol, the warning information is used to trigger the target vehicle in front of the vehicle to perform a warning operation in response to the fire event.

[0109] This embodiment can divide the device into functional modules based on the above method example. For example, each module can correspond to a separate function, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0110] When each functional module is divided according to its corresponding function, the device may further include an acquisition module, a prediction module, a determination module, and an output module. It should be noted that all relevant content regarding the steps involved in the above method embodiments can be referenced to the functional descriptions of the corresponding functional modules, and will not be repeated here.

[0111] It should be understood that the device provided in this embodiment is used to execute the above-described vehicle risk warning method, and therefore can achieve the same effect as the above-described implementation method.

[0112] When using an integrated unit, the device may include a processing module and a storage module. When the device is applied to a vehicle, the processing module can be used to control and manage the vehicle's movements. The storage module can be used to support the vehicle in executing program code, etc.

[0113] The processing module may be a processor or a controller, which can implement or execute various exemplary logic blocks, modules, and circuits as disclosed in this application. The processor may also be a combination of computing functions, such as a combination of one or more microprocessors, a combination of digital signal processing (DSP) and microprocessors, etc., and the storage module may be a memory.

[0114] In addition, the device provided in the embodiments of this application may specifically be a chip, component or module. The chip may include a connected processor and a memory. The memory is used to store instructions. When the processor calls and executes the instructions, the chip can execute a vehicle risk warning method provided in the above embodiments.

[0115] This application also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer, it causes the computer to execute the above-described related method steps to implement a vehicle risk warning method provided in the above embodiments. The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, Digital Video Discs (DVDs), Compact Disc Read-Only Memory (CD-ROMs), microdrives, and magneto-optical disks, read-only memory (ROMs), random access memory (RAMs), erasable programmable read-only memory (EPROMs), electrically erasable programmable read-only memory (EEPROMs), dynamic random access memory (DRAMs), video random access memory (VRAMs), flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of media or device suitable for storing instructions and / or data.

[0116] This application also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned steps to implement a vehicle risk warning method provided in the above embodiments.

[0117] The vehicle, computer-readable storage medium, computer program product or chip provided in this application are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can be referred to the beneficial effects of the corresponding methods provided above, and will not be repeated here.

[0118] Through the above description of the embodiments, those skilled in the art will 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.

[0119] In the 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 device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0120] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0121] The embodiments described above are merely illustrative of several implementation methods outlined in this specification. While the descriptions are specific and detailed, they should not be construed as limiting the scope of the solutions provided in this specification. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this specification, and these all fall within the scope of protection of this specification. Therefore, the scope of protection for this patent should be determined by the appended claims.

Claims

1. A method for risk warning of a vehicle, characterized in that, include: Acquire at least one type of perception information of a target vehicle in front of the vehicle, wherein the vehicle and the target vehicle in front of the vehicle are interconnected via a specified communication protocol, the specified communication protocol is configured with a message type identifier, and the message type identifier contains a field indicating warning information; Based on the various sensed information, determine whether a target vehicle in front of the vehicle has caught fire, so as to obtain event information; If the event information indicates that a fire has occurred in a target vehicle ahead of the vehicle, a warning message is sent to the target vehicle ahead of the vehicle through the specified communication protocol. The warning message is used to trigger the target vehicle ahead of the vehicle to perform a warning operation in response to the fire event.

2. The method according to claim 1, characterized in that, The acquisition of at least one type of perception information of a target vehicle ahead of the vehicle includes: Visual acquisition is performed based on the visual sensors configured in the vehicle to obtain image information about the target vehicle in front of the vehicle; Thermal imaging is performed based on the infrared sensors configured in the vehicle to obtain thermal information about the target vehicle in front of the vehicle. Particle density detection is performed based on the particulate matter sensor configured in the vehicle to obtain concentration information about the target vehicle in front of the vehicle; The image information, thermal information, and concentration information are aligned in time dimension by using timestamps to obtain perception information about the target vehicle in front of the vehicle.

3. The method according to claim 1 or 2, characterized in that, The step of determining whether a target vehicle in front of the vehicle has caught fire based on various sensed information, in order to obtain event information, includes: Feature extraction is performed based on the modal data distribution corresponding to the perceived information to obtain the perceived features, which include at least one of the following: image features of the flame, data distribution features in the thermal map, and data distribution features in the particle density map. The perceived features are respectively input into the corresponding attention branches for feature enhancement to obtain joint features, and the attention branches intersect with each other; Based on the combined features, a fire event is determined for the target vehicle in front of the vehicle to obtain the event information.

4. The method according to claim 3, characterized in that, The method further includes: The joint features are input into the convolutional layer associated with the attention branch to perform a full connection on the joint features to obtain global features; The global features are mapped onto image features to obtain a confidence map indicating the location of the fire; The confidence map is used to mark the location of the fire on the target vehicle in front of the vehicle.

5. The method according to claim 4, characterized in that, The step of marking the fire location of the target vehicle in front of the vehicle using the confidence map includes: Obtain the video stream associated with the confidence map; Based on the video stream, determine a plurality of adjacent reference video frames of the confidence map; If the confidence map corresponding to the reference video frame meets the preset conditions, then the fire location of the target vehicle in front of the vehicle indicated by the confidence map is marked.

6. The method according to claim 1, characterized in that, If the event information indicates that a target vehicle ahead of the vehicle has caught fire, then a warning message is sent to the target vehicle ahead of the vehicle via the specified communication protocol, including: If the event information indicates that a target vehicle in front of the vehicle has caught fire, then the vehicle identifier corresponding to the target vehicle in front of the vehicle is determined. The intensity level of the fire event is determined based on the event information; Configure location parameters based on fire location markers on confidence maps; Configure action codes to indicate warning actions based on the location parameters and intensity levels; The warning information is obtained by configuring a message type identifier using the vehicle identification, intensity level, location parameters, and motion code. Warning information is sent to the target vehicle ahead of the vehicle through the specified communication protocol.

7. The method according to claim 6, characterized in that, The configuration of action coding based on the location parameters and the intensity level includes: Configure the first action by combining the location parameters and the intensity level; Obtain environmental information corresponding to the target vehicle in front of the vehicle; The first action is adjusted based on the environmental information to obtain the second action; The action code is configured according to the second action.

8. A method for risk warning of a vehicle, characterized in that, include: The system acquires warning information sent by a vehicle, wherein the vehicle is interconnected with a target vehicle in front of the vehicle via a specified communication protocol, and the warning information is used to indicate that a fire has occurred in the target vehicle in front of the vehicle. The fire event is determined based on at least one type of perception information of the target vehicle in front of the vehicle. The perception information is collected by a target sensor configured in the vehicle. The specified communication protocol is configured with a message type identifier, and the message type identifier contains a field indicating the warning information. In response to the receipt of the warning information, the processing action indicated by the warning information is triggered.

9. A vehicle, characterized in that, The vehicles include: Memory, used to store executable program code; A processor is configured to call and run the executable program code from the memory, causing the vehicle to perform the vehicle risk warning method as described in any one of claims 1 to 8.

10. An electronic device, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the vehicle risk warning method according to any one of claims 1 to 8.

Images