A vehicle safety warning method and device, a storage medium and an equipment

Abstract

The application provides a vehicle safety warning method and device, a storage medium and equipment. In the method, in response to the state of the vehicle being an off state, it is detected whether there is an item with a light focusing effect in the passenger compartment of the vehicle, and a detection result is obtained. If the detection result indicates that the item exists in the passenger compartment, based on the position information of the item and the environmental information of the vehicle, the fire risk level of the vehicle is determined, and a warning operation corresponding to the fire risk level is performed. In this way, the risk of fire caused by the focusing of light rays under external light irradiation of the items left in the vehicle to form a high-temperature focal point is effectively avoided, thereby improving the safety of the vehicle.

Description

Technical Field

[0001] This application relates to the field of vehicle safety technology, and more specifically, to a vehicle safety early warning method, device, storage medium, and equipment. Background Technology

[0002] With the rapid development of the automotive industry and the continuous growth of car ownership, vehicle safety has become a focus of public concern. Among these concerns, the safety risks posed by items left inside vehicles are increasingly being recognized. Currently, relevant technologies primarily monitor risks by detecting flammable and explosive items such as lighters and perfumes left inside the vehicle, while the fire risks posed by items like water cups and beverage bottles are often overlooked. Fire safety experiments and real-world cases show that sunlight refracted through transparent containers filled with water can create localized high-temperature focal points, igniting flammable materials such as seats and anti-slip mats, thus causing a fire. Existing vehicles lack early warning systems to address this type of risk, compromising vehicle safety. Summary of the Invention

[0003] The purpose of this application is to provide a vehicle safety early warning method, device, storage medium and equipment, aiming to solve the problem that there is a lack of early warning means for the risk of fire caused by the light-focusing effect of transparent water containers, such as transparent water cups and glass bottles, due to external light, which adversely affects vehicle safety.

[0004] In a first aspect, this application provides a vehicle safety early warning method, comprising: in response to the vehicle being in an off state, detecting whether a target item exists in the passenger compartment of the vehicle, and obtaining a detection result; the target item includes an item with a focusing effect; if the detection result indicates that the target item exists in the passenger compartment, determining the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle; and performing an early warning operation corresponding to the fire risk level.

[0005] In the above implementation process, in response to the vehicle being in an off state, the system detects whether there are any objects with a focusing effect in the passenger compartment. If the detection result indicates the presence of such an object, the fire risk level of the vehicle is determined based on the object's location information and the vehicle's environmental information, and a warning operation corresponding to that fire risk level is executed. This effectively avoids the risk of fire caused by objects left inside the vehicle focusing sunlight to form a high-temperature focal point, thereby improving vehicle safety.

[0006] Furthermore, in some examples, detecting whether a target item exists in the passenger compartment of the vehicle includes: acquiring image information and depth information of the passenger compartment; inputting the image information into a pre-trained target item recognition model to obtain an inference result from the target item recognition model representing whether the target item exists in the passenger compartment; extracting the edge contours of each item in the passenger compartment from the depth information, analyzing whether the target item exists in the passenger compartment based on the edge contours, and obtaining an analysis result; and determining whether the target item exists in the passenger compartment based on the inference result and the analysis result.

[0007] In the above implementation process, the presence of target items inside the vehicle is detected by combining in-vehicle images and depth information, thereby effectively reducing the false detection rate and achieving accurate detection.

[0008] Furthermore, in some examples, the target object recognition model integrates a YOLOv8 backbone network and a convolutional block attention module; the YOLOv8 backbone network is used to extract feature maps of multiple scales from the image information; the convolutional block attention module is embedded after at least one C2f module in the YOLOv8 backbone network or in the neck network, and is used to sequentially perform channel attention enhancement and spatial attention enhancement on the feature maps to obtain enhanced feature maps.

[0009] In the above implementation process, CBAM is embedded in YOLOv8 to build a target object recognition model. For the input image, the YOLOv8 backbone network extracts preliminary features, and CBAM performs attention weighting on the feature map in the channel dimension and spatial dimension, thereby enhancing the target object recognition model's ability to extract features from the specular reflection area and refractive distortion area of ​​transparent objects and improving the accuracy of model detection.

[0010] Furthermore, in some examples, determining whether the target item exists in the passenger compartment based on the reasoning result and the analysis result includes: determining that the target item exists in the passenger compartment when both the reasoning result and the analysis result indicate that the target item exists in the passenger compartment.

[0011] In the above implementation process, the existence of a target item is only determined when both visual features and depth features meet the conditions. This can effectively reduce the false detection rate.

[0012] Furthermore, in some examples, the environmental information of the vehicle includes light intensity; determining the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle includes: determining whether the target item is located on a flammable material surface and whether it is on the path of external light incidence based on the location information of the target item; and determining the fire risk level of the vehicle as Level 1 fire risk when the target item is located on a flammable material surface and on the path of external light incidence, and the light intensity is greater than a preset light intensity threshold.

[0013] In the above process, when the target item is on a flammable material surface and in the path of external light, if the light intensity is greater than the preset light intensity, it indicates that the target item is in a high-intensity light environment. At this time, the fire risk level of the current vehicle is determined to be Level 1, indicating a high risk of fire. In this way, accurate identification of fire risk inside the vehicle is achieved.

[0014] Furthermore, in some examples, the environmental information of the vehicle also includes the vehicle's parking location and weather conditions; determining the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle further includes: determining the fire risk level of the vehicle as a level two fire risk when the vehicle is parked in an open area, the weather conditions are cloudy or overcast, and the target item is on a flammable material surface; and determining the fire risk level of the vehicle as a level three fire risk when the vehicle is parked in an indoor parking lot or the target item is not on a flammable material surface.

[0015] In the above implementation process, a rule-based logical judgment method is adopted, which combines the location of objects, light intensity, vehicle parking location and weather conditions to identify the fire risk level of the vehicle, thus achieving accurate identification of the risk of fire inside the vehicle.

[0016] Furthermore, in some examples, determining the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle includes: inputting the location information of the target item and the environmental information of the vehicle into a risk assessment model to obtain a predicted fire probability value output by the risk assessment model; if the predicted fire probability value is greater than a first threshold, determining the fire risk level of the vehicle as Level 1 fire risk; if the predicted fire probability value is less than or equal to the first threshold and greater than a second threshold, determining the fire risk level of the vehicle as Level 2 fire risk; if the predicted fire probability value is less than or equal to the second threshold, determining the fire risk level of the vehicle as Level 3 fire risk.

[0017] In the above implementation process, the accuracy of vehicle fire risk assessment is improved based on machine learning models.

[0018] Furthermore, in some examples, the execution of the early warning operation corresponding to the fire risk level includes: in the case of a level 1 fire risk, executing an in-vehicle audible and visual alarm and pushing a first early warning message to the user terminal; in the case of a level 2 fire risk, pushing a second early warning message to the user terminal; and in the case of a level 3 fire risk, recording a log.

[0019] In the above implementation process, the timeliness and reliability of early warnings are improved through tiered early warning systems, effectively enhancing vehicle safety and user experience.

[0020] Secondly, this application provides a vehicle safety warning device, comprising: a detection module, configured to detect whether a target item exists in the passenger compartment of the vehicle in response to the vehicle being in an off state, and obtain a detection result; the target item includes an item with a focusing effect; a determination module, configured to determine the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle if the detection result indicates that the target item exists in the passenger compartment; and a warning module, configured to perform a warning operation corresponding to the fire risk level.

[0021] Thirdly, this application provides an electronic device comprising: 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 steps of the method described in any of the first aspects.

[0022] Fourthly, this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described in any of the first aspects.

[0023] Other features and advantages disclosed in this application will be set forth in the following description, or some features and advantages may be inferred from the description or determined without doubt, or may be learned by practicing the above-described technology disclosed in this application.

[0024] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0025] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A flowchart illustrating a vehicle safety warning method provided in this application embodiment; Figure 2 A schematic diagram illustrating the workflow of a solution for preventing the risk of objects from focusing inside a vehicle, provided in an embodiment of this application; Figure 3 A block diagram of a vehicle safety warning device provided in an embodiment of this application; Figure 4 This is a structural block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0027] The technical solutions in the embodiments of this application will now be described with reference to the accompanying drawings.

[0028] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0029] Vehicle safety is a key concern for consumers, with the safety risks posed by items left inside vehicles increasingly gaining attention. Currently, detection technologies for items left in vehicles mainly focus on three areas: first, general item detection, which uses image differential and millimeter-wave radar to determine if items are left inside and alerts the owner; second, hazardous material detection, which focuses on monitoring the risks of flammable and explosive items such as lighters and perfumes; and third, live animal detection, which identifies and alerts when children or pets are left in the vehicle. However, these technologies do not address another type of safety hazard in vehicles—the risk of fire caused by sunlight focusing on transparent water containers, such as transparent cups and glass bottles. Firefighting experiments and real-world cases show that sunlight refracted through transparent water containers can create localized high-temperature focal points, igniting flammable materials such as seats and anti-slip mats, leading to fires. Existing vehicles lack warning mechanisms for this type of risk, thus negatively impacting vehicle safety.

[0030] To address the aforementioned issues, this application provides a vehicle safety early warning method. Responding to a vehicle in an off-state, the method detects the presence of objects with a focusing effect in the passenger compartment. If the detection result indicates the presence of such an object, the method determines the vehicle's fire risk level based on the object's location information and the vehicle's environmental information, and executes an early warning operation corresponding to that level. This effectively avoids the risk of fire caused by objects left inside the vehicle focusing sunlight to form a high-temperature focal point, thereby improving vehicle safety.

[0031] The embodiments of this application will be described below: like Figure 1 As shown, Figure 1 This is a flowchart of a vehicle safety warning method provided in an embodiment of this application. The method includes: Step 101: In response to the vehicle being in an off state, detect whether there is a target item in the passenger compartment of the vehicle and obtain the detection result; the target item includes items with a focusing effect. The target object mentioned in this step can refer to an object that can focus incident sunlight to a focal point through refraction or reflection, such as a plastic bottle filled with water, a glass, a magnifying glass, or a crystal ball. In this embodiment, when the user turns off the engine and gets out of the vehicle, onboard sensors collect data from inside the vehicle. The vehicle's domain controller or cloud server then analyzes the data to determine whether there are any objects with a light-focusing effect in the passenger compartment, thereby assessing whether there are any safety hazards in the vehicle.

[0032] In some embodiments, detecting whether a target item exists in the passenger compartment of the vehicle mentioned in this step may include: acquiring image information and depth information of the passenger compartment; inputting the image information into a pre-trained target item recognition model to obtain an inference result from the target item recognition model representing whether the target item exists in the passenger compartment; extracting the edge contours of each item in the passenger compartment from the depth information, analyzing whether the target item exists in the passenger compartment based on the edge contours, and obtaining an analysis result; and determining whether the target item exists in the passenger compartment based on the inference result and the analysis result. In other words, RGB (red, green, and blue) cameras inside the passenger cabin can capture image information, which is then input into a pre-trained target object recognition model. The model extracts visual features from the images and detects the presence of target objects based on these features. Simultaneously, depth information from a depth camera or structured light sensor is acquired, and a 3D point cloud is generated from the depth map. This point cloud is then processed using a point cloud processing network, such as PointNet++ (a deep learning network for processing point cloud data), or edge detection is performed directly on the depth map to obtain the 3D (three-dimensional) contours of each object in the passenger cabin. The extracted 3D contours are then aligned with and similarity calculated to a pre-stored 3D template of the target object, thus analyzing the presence of the target object. After completing visual and geometric detection, the results of both methods are combined to determine the presence of the target object, effectively reducing the false detection rate and achieving accurate detection.

[0033] Furthermore, in some embodiments, the aforementioned target object recognition model can integrate a YOLOv8 backbone network and a convolutional block attention module; wherein, the YOLOv8 backbone network is used to extract feature maps of multiple scales from the image information; the convolutional block attention module is embedded after at least one C2f module in the YOLOv8 backbone network or in the neck network, and is used to sequentially perform channel attention enhancement and spatial attention enhancement on the feature maps to obtain enhanced feature maps. That is to say, the target object recognition model used in this application can be obtained by introducing CBAM (Convolutional Block Attention Module) into YOLOv8 (a real-time target detection model). Taking a transparent water bottle in a car as an example, the standard YOLOv8 model "sees" more of the seat texture and environmental reflection on the bottle, so it is easy to misidentify the transparent water bottle as part of the background, or to miss it due to weak features. Based on this, embodiments of this application embed CBAM after at least one C2f module (an efficient feature extraction module) of YOLOv8 or in the neck network to construct a target object recognition model. Thus, when an in-vehicle image is input into the target object recognition model, the YOLOv8 backbone network extracts feature maps at multiple scales from the in-vehicle image. CBAM performs channel and spatial attention weighting on the individual feature maps output by the C2f module, or, in the neck network, performs channel and spatial attention weighting on the feature maps at each scale, thereby obtaining enhanced feature maps. These enhanced feature maps are then fed into the detection head. At this point, the model no longer "sees" a blurry background, but rather a clearly defined and feature-rich object. In this way, the target object recognition model can achieve accurate detection.

[0034] In its implementation, CBAM includes a channel attention submodule and a spatial attention submodule. The channel attention submodule performs global average pooling and max pooling on the input feature map, respectively. The pooling results are then fed into a shared multilayer perceptron (MLP). The outputs of the MLP are summed and passed through a sigmoid activation function (a non-linear activation function) to generate channel attention weights. These channel attention weights are multiplied by the input feature map to obtain the channel-enhanced feature map. The spatial attention submodule performs average pooling and max pooling on the channel-enhanced feature map along the channel dimension, respectively. The pooling results are concatenated and passed through a convolutional layer and a sigmoid activation function to generate spatial attention weights. These spatial attention weights are multiplied by the channel-enhanced feature map to output the attention-enhanced feature map, which is the final enhanced feature map. Taking a transparent water bottle inside a car as an example, in the channel attention submodule, based on max pooling, the model can capture the most prominent features in the image, such as highlights on the bottle or sharp edges caused by refraction. Through MLP calculation, the model automatically assigns higher weights to channels containing highlight or edge information and lowers the weights to channels containing ordinary background textures. The spatial attention submodule can capture the highlight reflection area and refractive distortion area of ​​the transparent water bottle. Specifically, because the highlight reflection area of ​​the transparent water bottle is extremely bright, it will stand out in the max pooling image, resulting in a high attention weight. The background texture behind the bottle is distorted, causing drastic changes in local gradients, which will also be captured by the spatial attention submodule. As for the flat seat area, its weight will be suppressed to a very low level. Therefore, based on CBAM, the target object recognition model can ignore background interference and focus on the two physical features of the transparent object: the highlight reflection area and the refractive distortion area, thereby achieving accurate detection.

[0035] In some embodiments, the training process of the aforementioned target object recognition model may include: acquiring a training dataset; the training dataset containing multiple image samples labeled with target object tags; constructing an initial target object recognition model; and training the initial target object recognition model using the training dataset with the goal of minimizing loss, to obtain a trained target object recognition model. That is, when training the target object recognition model, multiple image samples can be collected first, such as constructing a dataset containing water bottles inside a car with different materials, shapes, and liquid volumes, and then augmenting it, followed by manually adding labels to form a training dataset; then, an initial target object recognition model is constructed, which can be a YOLOv8 backbone network with a Convolutional Block Attention (CBAM) module embedded within it; next, the target object recognition model is trained using the training dataset, substituting the model's prediction results and the true labels into the loss function to obtain a loss value. By automatically adjusting the model's internal parameters, the loss value is minimized as much as possible. When the loss value is reduced to an acceptable range, or the number of iterations reaches a preset upper limit, the model training is complete. Thus, a target object recognition model capable of accurately detecting the presence of target objects in the passenger compartment based on in-vehicle images can be obtained.

[0036] Furthermore, in some embodiments, determining whether the target item exists in the passenger compartment based on the inference result and the analysis result, as mentioned above, may include: determining that the target item exists in the passenger compartment when both the inference result and the analysis result indicate that the target item exists in the passenger compartment. That is, the presence of the target item is only determined when both visual and depth features simultaneously meet the conditions, thus effectively reducing the false detection rate.

[0037] Step 102: If the detection result indicates that the target item exists in the passenger compartment, determine the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle; The location information of the target object mentioned in this step refers to its 3D coordinates in the vehicle coordinate system. This coordinates can be calculated using camera intrinsic parameters (parameters related to the camera's characteristics, including focal length, principal point coordinates, and distortion coefficients) and depth information. When an object with a focusing effect is detected inside the vehicle, its location information, combined with the vehicle's environmental information, is used to conduct a risk assessment, thereby achieving proactive safety protection and improving vehicle safety.

[0038] In some embodiments, the environmental information of the vehicle mentioned in this step may include light intensity; accordingly, determining the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle mentioned in this step may include: determining whether the target item is located on a flammable material surface and whether it is on the path of external light incidence based on the location information of the target item; if the target item is located on a flammable material surface and on the path of external light incidence, and the light intensity is greater than a preset light intensity threshold, determining the fire risk level of the vehicle to be Level 1 fire risk. That is, using the location result of the target item, determining whether the target item is located on a flammable material surface such as a dashboard or seat, and combining the angle of external light incidence and the vehicle's orientation, determining whether the target item is on the path of external light incidence. When a target object is on a flammable material surface and in the path of external light, if the light intensity exceeds a preset intensity, such as 50,000 lux, meaning the target object is in a high-intensity light environment, the object will concentrate the external light energy into a small area. The surfaces of flammable materials such as dashboards and seats will absorb heat and their temperature will rise rapidly. Once the temperature exceeds the ignition point, a fire can easily occur. Therefore, the current vehicle's fire risk level is determined to be Level 1, indicating a high risk of fire. This achieves accurate identification of fire risks inside the vehicle. The external light can be sunlight or artificial light sources, such as spotlights or searchlights; the light intensity can be detected by a light sensor installed on the vehicle. The preset light intensity is set by the user according to actual needs, and this application does not further limit it.

[0039] Furthermore, the environmental information of the vehicle mentioned in this step can also include the vehicle's parking location and weather conditions. Accordingly, determining the fire risk level of the vehicle based on the location information of the target item and the vehicle's environmental information can also include: determining the vehicle's fire risk level as Level 2 if the vehicle is parked in an open area, the weather is cloudy or overcast, and the target item is on a flammable material surface; and determining the vehicle's fire risk level as Level 3 if the vehicle is parked in an indoor parking lot or the target item is not on a flammable material surface. In other words, in addition to the item's location and light intensity, the vehicle's fire risk level can also be identified by combining the vehicle's parking location and weather conditions. Specifically, when the target item is on a flammable material surface, the vehicle is parked in an open area, and the weather is cloudy or overcast, although the light intensity is not greater than the preset light intensity threshold, the target item may still cause the flammable materials inside the vehicle to ignite over a prolonged period. Therefore, the current fire risk level of the vehicle is determined to be Level 2, indicating a potential fire risk. When the vehicle is parked in an indoor parking lot, since indoor parking lots generally do not receive direct sunlight, the artificial light intensity used is usually low, and there is virtually no risk of fire caused by focused light from transparent water bottles, magnifying glasses, etc. Therefore, the current fire risk level of the vehicle is determined to be Level 3, indicating an extremely low fire risk. Alternatively, when the target item is not on a flammable material surface, such as when it is placed in the glove box, there is also virtually no risk of fire caused by focused light from transparent water bottles, magnifying glasses, etc. Therefore, the current fire risk level of the vehicle is determined to be Level 3. This allows for accurate identification of fire risks inside the vehicle.

[0040] Furthermore, in some embodiments, determining the fire risk level of a vehicle based on the location information of the target item and the environmental information of the vehicle, as mentioned in this step, may include: inputting the location information of the target item and the environmental information of the vehicle into a risk assessment model to obtain a predicted fire probability value output by the risk assessment model; if the predicted fire probability value is greater than a first threshold, determining the fire risk level of the vehicle as Level 1 fire risk; if the predicted fire probability value is less than or equal to the first threshold and greater than a second threshold, determining the fire risk level of the vehicle as Level 2 fire risk; if the predicted fire probability value is less than or equal to the second threshold, determining the fire risk level of the vehicle as Level 3 fire risk. In other words, in addition to rule-based logical judgment, machine learning models can also be used to implement vehicle fire risk assessment. Specifically, the 3D coordinates of the target object within the vehicle's coordinate system, combined with information such as the vehicle's parking location, orientation, angle of incidence of external light, light intensity, predicted duration of vehicle stay, and weather conditions, are input into the risk assessment model. The model analyzes the current vehicle scenario based on this input information to predict the probability of a fire. If the predicted fire probability is greater than a first threshold, indicating a high probability of fire, the vehicle is classified as a Level 1 fire risk. If the predicted fire probability is less than or equal to the first threshold but greater than a second threshold, indicating a low probability of fire but still a potential risk, the vehicle is classified as a Level 2 fire risk. If the predicted fire probability is less than or equal to the second threshold, indicating an extremely low risk, the vehicle is classified as a Level 3 fire risk. This machine learning model improves the accuracy of in-vehicle fire risk assessment. The risk assessment model can employ any of the following machine learning models: random forest, logistic regression, or XGBoost (eXtreme Gradient Boosting). Its training dataset includes multiple samples, each containing item location information, vehicle environmental information, and a corresponding risk label. For details on the model training process, please refer to the relevant technical descriptions of the corresponding algorithms; this application will not elaborate further. Furthermore, the first and second thresholds can be set according to the specific needs of the scenario; this application does not impose any restrictions on this.

[0041] Step 103: Execute the early warning operation corresponding to the fire risk level.

[0042] In this embodiment, when the fire risk level of a vehicle is identified, a corresponding warning operation is performed based on the fire risk level. Different fire risk levels correspond to different warning operations, thus ensuring vehicle safety while improving the user experience.

[0043] In some embodiments, this step may include: in the case of a fire risk level of Level 1 fire risk, executing an in-vehicle audible and visual alarm and pushing a first warning message to the user terminal; in the case of a fire risk level of Level 2 fire risk, pushing a second warning message to the user terminal; and in the case of a fire risk level of Level 3 fire risk, recording a log. In other words, when the fire risk level is Level 1, meaning the vehicle has a high risk of fire, a visual alarm is triggered via warning lights on the dashboard and ambient lighting inside the vehicle. Simultaneously, a specific alert is emitted through devices such as a buzzer and car audio system to attract the attention of the driver and surrounding individuals. A strong reminder is also sent to the user via an app and / or SMS to ensure accurate information delivery, helping the user efficiently perceive and address any abnormal vehicle conditions promptly. When the fire risk level is Level 2, meaning the vehicle has a potential fire risk, a standard notification is sent to the user's device due to the longer warning time; the driver can handle the situation at their convenience. When the fire risk level is Level 3, meaning the fire risk is extremely low, data such as image information, depth information, vehicle environmental information, fire risk level, and timestamps are compiled into an event and written to the log without a notification, thus avoiding disturbance to the user. In this tiered warning system, the timeliness and reliability of warnings are improved, effectively enhancing vehicle safety and the user experience.

[0044] In this embodiment, when the vehicle is off, the system detects whether there are any objects with a focusing effect in the passenger compartment. If the detection result is yes, the system uses the location information of the object combined with the vehicle's environmental information to assess the vehicle's fire risk. Based on the identified fire risk level, the system executes corresponding warning actions. This effectively avoids the risk of fire caused by objects left inside the vehicle focusing sunlight to form a high-temperature focal point, thereby improving vehicle safety.

[0045] To provide a more detailed explanation of the solution in this application, a specific embodiment is described below: This embodiment provides a solution to prevent the risk of objects focusing on curved surfaces inside a vehicle. The system involved in this solution mainly consists of three parts: a perception layer, a decision layer, and an interaction layer. The perception layer is responsible for collecting images, depth information, and vehicle environmental data inside the vehicle; the decision layer runs the core algorithm to perform object recognition, localization, and risk logic judgment; and the interaction layer is responsible for converting the risk level into a user-perceptible reminder, such as a pop-up window on the vehicle's infotainment system, a push notification on the mobile phone, or flashing ambient lights.

[0046] Taking a transparent water bottle as an example of a curved surface item inside the car, the workflow of this solution is as follows: Figure 2 As shown, it includes: S201. Acquire in-vehicle images, depth information, and vehicle environment data; Specifically, this embodiment is triggered when the user turns off the engine, gets out of the car, and locks the vehicle. The in-vehicle image can be obtained by a high dynamic range RGB camera, such as an in-vehicle driver monitoring camera or a cabin panoramic camera; depth information can be obtained by a near-infrared depth camera or a structured light sensor; vehicle environmental data includes light intensity, vehicle parking position, weather conditions, etc., where light intensity can be obtained by a light sensor mounted on the vehicle; vehicle parking position can be obtained by GPS (Global Positioning System) combined with map matching; weather conditions can be obtained through a network weather API (Application Programming Interface); S202. Use a detection model to process the images inside the vehicle to detect whether there is a transparent water bottle inside the vehicle. Specifically, the detection model is an improved YOLOv8 model, which introduces CBAM into the YOLOv8 backbone network. CBAM is concatenated in the backbone or neck network of the YOLOv8 backbone network, and it recalibrates the feature map through channel and spatial dimensions. In implementation, the workflow of CBAM in YOLOv8 is as follows: First, the input is a scene captured by an in-vehicle camera, assuming the scene contains a water bottle placed on the center console; then, the YOLOv8 backbone network extracts preliminary features, at which point the bottle's features are mixed in with the background; then, CBAM intervenes. In the channel dimension, the model identifies channels containing highlight or edge information; in the spatial dimension, the model detects high brightness gradients on both sides of the bottle and texture transitions in the middle of the bottle. CBAM generates a mask, amplifying the weight of the area where the bottle is located and reducing the weight of the surrounding seats; then, the enhanced feature map is fed into the detection head; finally, the model accurately outlines the location of the water bottle. S203. Geometric detection based on depth information is used to detect whether there is a transparent water bottle inside the vehicle. Specifically, the depth information is processed using a point cloud processing network, or edge detection is performed directly on the depth map. Since transparent objects usually appear as "holes" or depth jumps in the depth map, morphological operations can be used to fill the "holes" and extract the contours during edge detection. Then, the extracted 3D contours are aligned and similarity is calculated with the 3D contours of various pre-stored transparent water bottles. The presence of transparent water bottles inside the vehicle is detected based on the similarity calculation results. S204. Do the visual features and depth features simultaneously meet the conditions? If yes, execute S205; otherwise, execute S208. Specifically, a post-fusion strategy is adopted, and the presence of a transparent water bottle in the vehicle is only determined when both the visual detection result and the geometric detection result indicate that a transparent water bottle exists in the vehicle, thereby reducing the false detection rate. S205. Calculate the 3D coordinates of the transparent water bottle in the vehicle coordinate system using camera intrinsic parameters and depth information; S206. Based on the 3D coordinates of the transparent water bottle and vehicle environmental data, determine the risk level; Specifically, the 3D coordinates of the transparent water bottle are used to determine whether it is located on the surface of flammable materials such as dashboards and seats, and whether it is on the path of sunlight. A vehicle is classified as Level 1 risk when it meets all of the following conditions: the vehicle is parked outdoors; the weather is sunny; the light intensity is greater than 50,000 lux and lasts for a certain period of time; a transparent water bottle is located on the surface of flammable materials such as the dashboard or seats; or the transparent water bottle is in the path of sunlight. A vehicle is classified as Level 2 risk when it meets all of the following conditions: the vehicle is parked outdoors; the weather is overcast or cloudy; and a transparent water bottle is located on the dashboard, seats, or other flammable material surfaces. A vehicle is classified as Level 3 risk if it meets any of the following conditions: the vehicle is parked in an indoor parking lot; the current time is night; or a transparent water bottle is located inside the glove box. S207. Based on the risk level, execute the corresponding early warning operation; Specifically, when the vehicle's risk level is Level 1, the system controls the central control display to output a red pop-up warning, accompanied by an urgent prompt sound. At the same time, a strong reminder is pushed through the APP: "A water bottle has been detected in the sunlight, posing a fire risk. Please handle it immediately." If the user does not take any action, a text message will be sent after 5 minutes. When the vehicle's risk level is Level 2, the system will send a regular notification via the app: "A water bottle has been detected inside the vehicle. We recommend that you store it properly." When the vehicle's risk level is Level 3, the system records the information silently without issuing a notification. S208, End of process.

[0047] This embodiment of the solution enables accurate detection of items left in the vehicle that have a focusing effect, thereby effectively avoiding the risk of fire caused by the high-temperature focus formed by the focused light from items left in the vehicle under sunlight, thus effectively ensuring vehicle safety. At the same time, it improves the user experience through graded warnings.

[0048] Corresponding to the embodiments of the aforementioned methods, this application also provides embodiments of a vehicle safety warning device and a terminal for its application: like Figure 3 As shown, Figure 3 This is a block diagram of a vehicle safety warning device provided in an embodiment of this application. The device includes: The detection module 31 is used to detect whether a target item exists in the passenger compartment of the vehicle in response to the vehicle being in an off state, and to obtain a detection result; the target item includes items with a focusing effect. The determination module 32 is used to determine the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle if the detection result indicates that the target item exists in the passenger compartment. The early warning module 33 is used to perform early warning operations corresponding to the fire risk level.

[0049] The specific implementation process of the functions and roles of each module in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0050] This application also provides an electronic device, please refer to [link to application]. Figure 4 , Figure 4 This is a structural block diagram of an electronic device provided in an embodiment of this application. The electronic device may include a processor 410, a communication interface 420, a memory 430, and at least one communication bus 440. The communication bus 440 is used to enable direct communication between these components. In this embodiment, the communication interface 420 of the electronic device is used for signaling or data communication with other node devices. The processor 410 may be an integrated circuit chip with signal processing capabilities.

[0051] The processor 410 described above can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor, or the processor 410 can be any conventional processor.

[0052] The memory 430 may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc. The memory 430 stores computer-readable instructions. When these computer-readable instructions are executed by the processor 410, the electronic device can perform the aforementioned operations. Figure 1 The various steps involved in the method implementation examples.

[0053] Alternatively, the electronic device may also include a storage controller and an input / output unit.

[0054] The memory 430, storage controller, processor 410, peripheral interface, and input / output unit are electrically connected directly or indirectly to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses 440. The processor 410 is used to execute executable modules stored in the memory 430, such as software function modules or computer programs included in electronic devices.

[0055] The input / output unit is used to provide users with the ability to create tasks and to set optional start periods or preset execution times for those tasks, thereby enabling user-server interaction. The input / output unit may be, but is not limited to, a mouse and keyboard.

[0056] Understandable. Figure 4 The structure shown is for illustrative purposes only; the electronic device may also include components that are more advanced than those shown. Figure 4 The more or fewer components shown, or having the same Figure 4 The different configurations shown. Figure 4 The components shown can be implemented using hardware, software, or a combination thereof.

[0057] This application also provides a storage medium storing instructions. When the instructions are run on a computer, the computer program is executed by a processor to implement the method described in the method embodiment. To avoid repetition, the method will not be described again here.

[0058] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of the apparatus and methods according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0059] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0060] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in 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, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0061] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application. It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0062] 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 technical scope 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.

[0063] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A vehicle safety early warning method, characterized in that, include: In response to the vehicle being in an off state, the system detects whether a target item exists in the passenger compartment of the vehicle and obtains a detection result; the target item includes items with a focusing effect. If the detection result indicates that the target item is present in the passenger compartment, the fire risk level of the vehicle is determined based on the location information of the target item and the environmental information of the vehicle. Perform the warning operation corresponding to the fire risk level.

2. The method according to claim 1, characterized in that, The detection of whether the target item exists in the passenger compartment of the vehicle includes: Acquire image information and depth information of the crew cabin; The image information is input into a pre-trained target object recognition model to obtain the inference result output by the target object recognition model representing whether the target object exists in the passenger compartment; and the edge contours of each object in the passenger compartment are extracted from the depth information, and the existence of the target object in the passenger compartment is analyzed based on the edge contours to obtain the analysis result. Based on the reasoning and analysis results, it is determined whether the target item exists in the crew cabin.

3. The method according to claim 2, characterized in that, in, The target object recognition model integrates the YOLOv8 backbone network and the convolutional block attention module; The YOLOv8 backbone network is used to extract feature maps of multiple scales from the image information; The convolutional block attention module is embedded after at least one C2f module in the YOLOv8 backbone network or in the neck network, and is used to sequentially perform channel attention enhancement and spatial attention enhancement on the feature map to obtain an enhanced feature map.

4. The method according to claim 2, characterized in that, The step of determining whether the target item exists in the crew cabin based on the reasoning result and the analysis result includes: If both the reasoning result and the analysis result indicate that the target item exists in the crew cabin, it is determined that the target item exists in the crew cabin.

5. The method according to claim 1, characterized in that, in, The vehicle's environmental information includes light intensity; The process of determining the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle includes: Based on the location information of the target item, it is determined whether the target item is located on a flammable material surface and whether it is on the path of external light incidence. If the target item is on a flammable material surface and in the path of external light, and the light intensity is greater than a preset light intensity threshold, the fire risk level of the vehicle is determined to be Level 1 fire risk.

6. The method according to claim 5, characterized in that, in, The vehicle's environmental information also includes the vehicle's parking location and weather conditions; The step of identifying the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle further includes: Given that the vehicle is parked in an open area, the weather conditions are cloudy or overcast, and the target item is on a flammable material surface, the fire risk level of the vehicle is determined to be Level 2 fire risk. If the vehicle is parked in an indoor parking lot, or if the target item is not on a flammable material surface, the fire risk level of the vehicle is determined to be Level 3 fire risk.

7. The method according to claim 1, characterized in that, The process of determining the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle includes: The location information of the target item and the environmental information of the vehicle are input into the risk assessment model to obtain the fire probability prediction value output by the risk assessment model. If the predicted fire probability value is greater than the first threshold, the fire risk level of the vehicle is determined to be Level 1 fire risk; if the predicted fire probability value is less than or equal to the first threshold and greater than the second threshold, the fire risk level of the vehicle is determined to be Level 2 fire risk; if the predicted fire probability value is less than or equal to the second threshold, the fire risk level of the vehicle is determined to be Level 3 fire risk.

8. The method according to claim 6 or 7, characterized in that, The execution of the early warning operation corresponding to the fire risk level includes: When the fire risk level is Level 1, the vehicle will activate an in-vehicle audible and visual alarm and push a first warning message to the user terminal. When the fire risk level is level two, a second early warning message is pushed to the user terminal; If the fire risk level is level three, log the information.

9. A vehicle safety warning device, characterized in that, include: The detection module is used to detect whether a target item exists in the passenger compartment of the vehicle in response to the vehicle being in an off state, and to obtain a detection result; the target item includes items with a focusing effect. The determination module is used to determine the fire risk level of the vehicle based on the location information of the target item and the environmental information of the vehicle if the detection result indicates that the target item exists in the passenger compartment. The early warning module is used to perform early warning operations corresponding to the fire risk level.

10. A computer-readable storage medium, characterized in that, It stores a computer program thereon, which, when executed by a processor, implements the method as described in any one of claims 1 to 8.

11. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in any one of claims 1 to 8.

Images