Weather identification method and electronic equipment

Abstract

The invention relates to the technical field of data processing, in particular to a weather identification method and electronic equipment. The method comprises the following steps: acquiring environment image data, environment audio data and environment perception data around a vehicle; inputting the environment image data, the environment audio data and the environment perception data into a predetermined rainfall probability determination model, and performing rainfall analysis to obtain a local rainfall probability of the vehicle; and sending the local rainfall probability to an edge device, so that the edge device performs regional rainfall identification based on the local rainfall probability. According to the method, the image data of the environment around the vehicle, the environment audio data and the environment sensing data are combined, accurate rainfall analysis is carried out by using the preset rainfall probability determination model, the obtained local rainfall probability is more accurate, and the vehicle end also sends the local rainfall probability to the edge equipment, so that the accuracy of rainfall analysis is improved. Regional rainfall recognition is carried out by utilizing edge equipment, and the accuracy of rainfall recognition is improved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a weather recognition method and electronic device. Background Technology

[0002] Currently, weather recognition for vehicles is generally determined based on detections from a single sensor on the vehicle (e.g., an image sensor).

[0003] Weather information identified by a single sensor is not accurate enough and can affect safe driving. Summary of the Invention

[0004] In view of this, the purpose of this application is to propose a weather recognition method and electronic device to solve the technical problem that the accuracy of weather information identified by using a single sensor in vehicles is relatively low.

[0005] To achieve the above objectives, this application provides a weather recognition method applied to a vehicle, the method comprising:

[0006] Acquire environmental image data, environmental audio data, and environmental perception data around the vehicle; The environmental image data, the environmental audio data, and the environmental perception data are input into a pre-determined rainfall probability determination model to perform rainfall analysis and obtain the local rainfall probability for the vehicle. The local rainfall probability is sent to the edge device so that the edge device can identify regional rainfall based on the local rainfall probability.

[0007] Based on the same inventive concept, this application also provides a weather recognition method applied to edge devices, the method comprising: Receive local rainfall probability data from multiple vehicle terminals within the area corresponding to the edge device; Receive local rainfall probability data from multiple vehicle terminals within the area corresponding to the edge device; Based on the local rainfall probability sent from each vehicle terminal, the characteristic data corresponding to the area range is determined; The feature data is input into a pre-trained regional rainfall recognition model to perform regional rainfall analysis and obtain regional rainfall recognition results. The rainfall identification results for the area are sent to each vehicle within the area.

[0008] Based on the same inventive concept, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor implements the method described above when executing the computer program.

[0009] As can be seen from the above, the weather recognition method and electronic device provided in this application can combine environmental image data, environmental audio data, and environmental perception data around the vehicle, and use a pre-set rainfall probability determination model to perform accurate rainfall analysis. This combines the three-dimensional data of images, audio, and environmental perception, making the local rainfall probability obtained by the rainfall probability determination model more accurate. Furthermore, the vehicle also sends the local rainfall probability to the edge device, allowing the edge device to combine the local rainfall probabilities sent by each vehicle. This breaks the information silo situation when relying on a single sensor for rainfall recognition, enabling more accurate regional rainfall recognition and ensuring higher accuracy of the regional rainfall recognition results. In this way, the vehicle can perform more accurate driving control by combining the local rainfall probability with the regional rainfall recognition results, thus improving driving safety. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram illustrating an application scenario of the weather recognition method according to an embodiment of this application; Figure 2 This is a flowchart of the weather recognition method executed on the vehicle side in an embodiment of this application; Figure 3 A logical schematic diagram of a weather recognition method performed by an edge device according to an embodiment of this application; Figure 4 This is a structural block diagram of a weather recognition device installed on a vehicle according to an embodiment of this application; Figure 5 This is a structural block diagram of a weather recognition device installed on an edge device according to an embodiment of this application; Figure 6 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.

[0013] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this application should have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms "first," "second," and similar terms used in the embodiments of this application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are only used to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0014] Definitions: ADAS: Advanced Driver Assistance Systems.

[0015] RSU: Road Side Unit (i.e., edge device).

[0016] T-BOX: Telematics BOX, a remote information processor.

[0017] LSTM: Long Short-Term Memory.

[0018] In related technologies, with the rapid development of intelligent vehicle connectivity technology, the demand for advanced driver assistance systems (ADAS) and autonomous driving functions to perceive adverse weather conditions is increasing. In adverse weather scenarios (e.g., rain), sensor performance is significantly reduced, affecting the road surface adhesion coefficient, which is a key factor threatening driving safety. Some weather recognition solutions in related technologies have the following drawbacks: One major drawback is the strong limitations of single-vehicle perception: many related technologies rely heavily on single-vehicle sensors, such as purely vision-based weather recognition methods or LiDAR point cloud analysis, which have low reliability in extreme weather or complex scenarios. For example, vision-based solutions perform poorly at night or in strong light, or LiDAR experiences increased noise in heavy rain.

[0019] The second defect is the isolation and lack of collaboration of data: In the relevant technologies, most of the weather recognition is performed by the vehicle itself, which is a "static" recognition. It lacks real-time data interaction between vehicles and between vehicles and roads, and cannot realize beyond-line-of-sight warning.

[0020] Thirdly, rigid data and poor adaptability: Most weather recognition solutions in related technologies are based on fixed thresholds for weather recognition and judgment, which cannot adapt to changing environments and have weak adaptability.

[0021] Fourthly, the weather recognition dimension is limited: related technologies can only identify a single weather condition and lack a refined assessment of multiple dimensions such as rainfall intensity, road slip risk, and safe following distance.

[0022] Based on the above description, the principles and spirit of this application will be explained in detail below with reference to several representative embodiments.

[0023] refer to Figure 1 This diagram illustrates an application scenario of the weather recognition method provided in this application. The application scenario includes a vehicle-side device 101 and an edge device 102. Both the vehicle-side device 101 and the edge device 102 can be connected via wired or wireless communication networks. The edge device 102 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network) architecture, and big data and artificial intelligence platforms.

[0024] Vehicle-mounted terminal 101 can acquire environmental image data, environmental audio data, and environmental perception data around the vehicle. It combines these data with a pre-set rainfall probability determination model to perform accurate rainfall analysis and obtain an accurate local rainfall probability. Vehicle-mounted terminal 101 then sends this local rainfall probability to edge device 102. Edge device 102 can then combine the local rainfall probabilities from multiple vehicle-mounted terminals 101 within its assigned area to determine the corresponding feature data. This feature data is then input into a pre-trained regional rainfall recognition model for accurate regional rainfall analysis, resulting in more accurate regional rainfall recognition results. Edge device 102 also sends these regional rainfall recognition results to each vehicle-mounted terminal 101 within its assigned area. This allows vehicle-mounted terminal 101 to combine local rainfall probabilities with regional rainfall recognition results for more accurate driving control, improving the safety of vehicle-mounted driving control.

[0025] The following is combined Figure 1The application scenarios described above illustrate the weather recognition method according to exemplary embodiments of this application. It should be noted that the above application scenarios are merely shown to facilitate understanding of the spirit and principles of this application, and the embodiments of this application are not limited in any way. Rather, the embodiments of this application can be applied to any applicable scenario.

[0026] Based on the above, the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0027] The weather recognition method proposed in the embodiments of this application is applied to a vehicle-mounted terminal device, which is a terminal device installed on a vehicle for data analysis.

[0028] like Figure 2 As shown, the method includes: Step 201: Acquire environmental image data, environmental audio data, and environmental perception data around the vehicle.

[0029] In specific implementation, environmental image data refers to images of the environment surrounding the vehicle captured by cameras installed on the vehicle. These cameras include at least one of the following: a front-facing camera, a rear-facing camera, a side-facing camera, and a panoramic camera. Environmental image data includes at least one of the following: a front-facing image of the vehicle taken by the front-facing camera, a rear-facing image of the vehicle taken by the rear-facing camera, a side-facing image of the vehicle taken by the side-facing camera, and a panoramic image of the vehicle taken by the panoramic camera.

[0030] Ambient audio data refers to audio data collected by a recording device installed on the vehicle, capturing the audio around the vehicle. This recording device is positioned on the exterior of the vehicle to collect audio data from its surroundings.

[0031] Environmental perception data refers to the perception data of the vehicle's surroundings collected by environmental perception sensors installed on the vehicle. These environmental perception sensors include at least one of the following: a temperature sensor, a humidity sensor, an atmospheric pressure sensor, and an inertial sensor. Environmental perception data includes at least one of the following: temperature information collected by the temperature sensor, humidity information collected by the humidity sensor, atmospheric pressure values ​​collected by the atmospheric pressure sensor, and vehicle attitude information collected by the inertial sensor.

[0032] In addition, to ensure the accuracy of environmental image data, environmental audio data, and environmental perception data, preprocessing will be performed on all environmental image data, environmental audio data, and environmental perception data. The preprocessing methods include at least one of the following: temporal alignment processing, filtering processing, and key feature extraction processing.

[0033] Step 202: Input the environmental image data, the environmental audio data, and the environmental perception data into a pre-determined rainfall probability determination model to perform rainfall analysis and obtain the local rainfall probability of the vehicle.

[0034] In practice, a rainfall probability determination model is pre-set on the vehicle. This model has a built-in rainfall probability algorithm. The environmental image data, environmental audio data, and environmental perception data obtained above are input into the rainfall probability determination model. The rainfall probability algorithm is used to perform rainfall analysis and processing to obtain the local rainfall probability. This can combine the three-dimensional data of images, audio, and environmental perception to perform rainfall analysis from multiple dimensions, making the obtained local rainfall probability more accurate.

[0035] Step 203: Send the local rainfall probability to the edge device so that the edge device can identify regional rainfall based on the local rainfall probability.

[0036] In practice, the vehicle-mounted terminal sends the obtained local rainfall probability to the edge device. In this way, the edge device can receive local rainfall probabilities from multiple vehicle-mounted terminals within its assigned area. The edge device combines the local rainfall probabilities from each vehicle-mounted terminal, thereby ensuring higher accuracy in the obtained regional rainfall identification results.

[0037] Then, the vehicle can receive the regional rainfall recognition results from the edge device, and correct the local rainfall probability based on the regional rainfall recognition results, so that the accuracy of the corrected local rainfall probability is higher, and then the vehicle can be controlled more safely based on the corrected local rainfall probability.

[0038] The above solution combines environmental image data, environmental audio data, and environmental perception data around the vehicle. A pre-defined rainfall probability determination model is then used for accurate rainfall analysis. This integrates three-dimensional data from images, audio, and environmental perception, making the local rainfall probability obtained by the rainfall probability determination model more accurate. Furthermore, the vehicle sends the local rainfall probability to edge devices, allowing these devices to combine the local rainfall probabilities from various vehicle terminals. This overcomes the information silos inherent in rainfall identification based on a single vehicle sensor, enabling more accurate regional rainfall identification and ensuring higher accuracy of the results. By combining the local rainfall probability with the regional rainfall identification results, the vehicle can perform more precise driving control, improving driving safety.

[0039] In some embodiments, after inputting the environmental image data, the environmental audio data, and the environmental perception data into a pre-determined rainfall probability determination model in step 202, the execution process of the rainfall probability determination model is as follows: Step 2021: Perform rainfall analysis based on the environmental image data to determine the first rainfall probability.

[0040] In practice, the rainfall probability determination model includes an environmental image rainfall analysis unit, which can perform rainfall analysis on the environmental image data fed back by the camera, and then determine the first rainfall probability corresponding to the environmental image data.

[0041] Step 2022: Perform rainfall analysis based on the environmental audio data to determine a second rainfall probability (e.g., P1).

[0042] In practice, the rainfall probability determination model includes an environmental audio rainfall analysis unit, which can perform rainfall analysis on the environmental audio data fed back by the audio equipment, and then determine the second rainfall probability (e.g., P2) corresponding to the environmental audio data.

[0043] Step 2023: Perform rainfall analysis based on the environmental perception data to determine the third rainfall probability.

[0044] In practice, the rainfall probability determination model includes an environmental sensing rainfall analysis unit, which can perform rainfall analysis on the environmental sensing data fed back by the environmental sensing sensor, and then determine the third rainfall probability (e.g., P3) corresponding to the environmental sensing data.

[0045] Step 2024: The first rainfall probability, the second rainfall probability, and the third rainfall probability are weighted and summed to obtain the local rainfall probability of the vehicle.

[0046] In practice, the weight values ​​corresponding to the first, second, and third rainfall probabilities are determined, and then the first, second, and third rainfall probabilities are weighted and summed according to their respective weight values. This results in a more accurate local rainfall probability.

[0047] Furthermore, the weight values ​​corresponding to the first, second, and third rainfall probabilities mentioned above can be adjusted and determined according to the actual situation, ensuring that the sum of the three weight values ​​equals 1.

[0048] The above scheme determines the first rainfall probability of environmental image data after rainfall analysis, the second rainfall probability of environmental audio data after rainfall analysis, and the third rainfall probability of environmental perception data after rainfall analysis. The three rainfall probabilities are then weighted and summed to obtain the local rainfall probability. This approach combines the three dimensions of image, audio, and environmental perception data, making the local rainfall probability obtained by the rainfall probability determination model more accurate.

[0049] In some embodiments, step 2024 includes: Step 20241: Obtain the image confidence, audio confidence, and environmental perception confidence of the rainfall probability determination model output at the previous moment.

[0050] In practical implementation, this rainfall probability determination model can also output the confidence scores corresponding to environmental image data, environmental audio data, and environmental perception data. This allows us to obtain the previous moment's image confidence score (corresponding to the environmental image data), audio confidence score (corresponding to the environmental audio data), and environmental perception confidence score (corresponding to the environmental perception data) output by the rainfall probability determination model when processing the previous moment's environmental image data, environmental audio data, and environmental perception data.

[0051] Since the model does not yet have the output results from the previous time step because the probability of rainfall is determined at the initial time step, the image confidence, audio confidence, and environmental perception confidence from the previous time step are all their respective initial values.

[0052] Step 20242: Determine the image weight value of the environmental image data based on the image confidence level of the previous time step.

[0053] In practice, the image weight value of the environmental image data is obtained by dividing the image confidence of the previous moment by the sum of the image confidence, audio confidence, and environmental perception confidence of the previous moment.

[0054] Step 20243: Determine the audio weight value of the environmental audio data based on the audio confidence level of the previous moment.

[0055] In practice, the audio confidence score of the previous moment is divided by the sum of the image confidence score, the audio confidence score, and the environmental perception confidence score of the previous moment to obtain the audio weight value of the environmental audio data.

[0056] Step 20244: Determine the perception weight value of the environmental perception data based on the environmental perception confidence level of the previous time step.

[0057] In practice, the perception weight value of the environmental perception data is obtained by dividing the previous moment's environmental perception confidence by the sum of the previous moment's image confidence, audio confidence, and environmental perception confidence.

[0058] Step 20245: Multiply the image weight value by the first rainfall probability to obtain a first product, multiply the audio weight value by the second rainfall probability to obtain a second product, multiply the perception weight value by the third rainfall probability to obtain a third product, and add the first product, the second product and the third product to obtain the local rainfall probability of the vehicle.

[0059] In practice, the probability of local rainfall The calculation formula is: ; in, Represents the image weight value. Indicates the audio weight value. Indicates the perceived weight value. Indicates the first probability of rainfall. Indicates the second probability of rainfall. This indicates the probability of the third rainfall.

[0060] Step 20246: Determine the image confidence level at the current moment based on the environmental image data, determine the audio confidence level at the current moment based on the environmental audio data, and determine the environmental perception confidence level at the current moment based on the environmental perception data.

[0061] In practice, while performing rainfall analysis to determine the local rainfall probability, the rainfall probability determination model also determines the image confidence level at the current moment based on the ambient light intensity in the environmental image data. If the ambient light intensity is lower than the minimum threshold or higher than the maximum threshold, the image confidence level at the current moment will be reduced to below a predetermined value. If the minimum threshold ≤ ambient light intensity ≤ optimal intensity value, the image confidence level at the current moment is positively correlated with the ambient light intensity. If the optimal intensity < ambient light intensity ≤ maximum threshold, the image confidence level at the current moment is negatively correlated with the ambient light intensity.

[0062] While performing rainfall analysis to determine the local rainfall probability, the rainfall probability determination model also determines the audio confidence level at the current moment based on the background noise in the environmental audio data. The audio confidence level at the current moment is negatively correlated with the decibel value of the background noise.

[0063] While performing rainfall analysis to determine the local rainfall probability, the rainfall probability determination model also determines the current environmental perception confidence level based on the health status of environmental perception sensors in the environmental perception data (a higher health status value indicates more accurate detection by the environmental perception sensors). This current environmental perception confidence level is positively correlated with the health status value. For example, when the environmental perception sensors are operating normally, the current environmental perception confidence level is maintained in the range of 0.8 to 1.

[0064] The values ​​of the current moment image confidence, the current moment audio confidence, and the current moment environmental perception confidence are all in the range of 0 to 1.

[0065] Step 20247: The rainfall probability determination model outputs the current moment image confidence, the current moment audio confidence, and the current moment environmental perception confidence, together with the local rainfall probability.

[0066] In practice, the rainfall probability determination model outputs the current image confidence level, the current audio confidence level, and the current environmental perception confidence level, along with the local rainfall probability.

[0067] In addition, this rainfall probability determination model can also perform data quality analysis on environmental image data, environmental audio data, and environmental perception data to determine the data quality of the environmental image data, environmental audio data, and environmental perception data currently obtained by the vehicle. This data quality is then output along with the aforementioned "current-time image confidence score, current-time audio confidence score, and current-time environmental perception confidence score, and local rainfall probability." The data quality value ranges from 0 to 1; higher data quality indicates higher accuracy of the environmental image data, environmental audio data, and environmental perception data obtained by the vehicle, and a more accurate determination of the local rainfall probability.

[0068] Through the above scheme, the rainfall probability determination model can determine the corresponding accurate image weight value, audio weight value, and perception weight value based on the image confidence value, audio confidence value, and environmental perception confidence value of the previous moment, respectively. This allows for the weighted summation of the first, second, and third rainfall probabilities to obtain the accurate local rainfall probability. Furthermore, it determines the confidence values ​​corresponding to the environmental image data, environmental audio data, and environmental perception data, facilitating the determination of the local rainfall probability in subsequent moments. This ensures that the rainfall probability determination model can continuously determine the local rainfall probability in chronological order.

[0069] In a preferred embodiment, the vehicle-mounted device also connects to a cloud server, which continuously updates and trains the rainfall probability determination model. The specific update and training process is as follows: (1) Collect sample environment image data, sample environment audio data and sample environment perception data from each vehicle end, and label the real local rainfall probability, real image confidence, real audio confidence and real environment perception confidence, and real data quality to form the first training sample. Multiple first training samples are combined to form the first training sample set.

[0070] (2) Input the first training samples in the first training sample set into the rainfall probability determination model in sequence. The rainfall probability determination model has an environmental image data analysis layer, which is used to perform rainfall analysis on the sample environmental image data in the first training sample to determine the training first rainfall probability.

[0071] (3) The rainfall probability determination model has an environmental audio data analysis layer, which is used to perform rainfall analysis on the sample environmental audio data in the first training sample to determine the training second rainfall probability.

[0072] (4) The rainfall probability determination model has an environmental perception data analysis layer, which is used to perform rainfall analysis on the environmental perception data of the first training sample to determine the rainfall probability of the third training sample.

[0073] (5) The rainfall probability determination model has an attention layer, which is used to weight and sum the training first rainfall probability, training second rainfall probability and training third rainfall probability to obtain the training local rainfall probability of the vehicle.

[0074] (6) The rainfall probability determination model has a confidence analysis layer, which is used to analyze the sample environment image data, sample environment audio data and sample environment perception data respectively, and determine the corresponding training image confidence, training audio confidence and training environment perception confidence respectively.

[0075] (7) The rainfall probability determination model has a data quality analysis layer, which is used to perform quality analysis on sample environment image data, sample environment audio data and sample environment perception data to obtain the training data quality.

[0076] (8) The rainfall probability determination model will be trained with local rainfall probability, as well as training image confidence, training audio confidence and training environment awareness confidence, and training data quality output.

[0077] (9) Compare the training local rainfall probability with the actual local rainfall probability, compare the training image confidence, training audio confidence and training environmental perception confidence with the actual image confidence, actual audio confidence and actual environmental perception confidence respectively, compare the training data quality with the actual data quality, determine the loss function based on each comparison result, calculate the corresponding loss value, update and optimize the parameters of each layer of the rainfall probability determination model based on the loss value, and obtain the optimized rainfall probability determination model.

[0078] (10) For each first training sample in the first training sample set, repeat the process from (2) to (9) based on the previously optimized rainfall probability determination model until the first training sample set is trained, or the final optimized rainfall probability determination model reaches the predetermined convergence level, or the accuracy of the final optimized rainfall probability determination model reaches the predetermined accuracy threshold, then stop training and use the final optimized rainfall probability determination model as the final optimized rainfall probability determination model.

[0079] (11) Send the final optimized rainfall probability determination model to the vehicle terminal so that the vehicle terminal can replace the original rainfall probability determination model with the final optimized rainfall probability determination model for rainfall analysis.

[0080] Based on the same inventive concept, a weather recognition method is applied to an edge device, which is a server set up on the roadside.

[0081] like Figure 3 As shown, the method includes: Step 301: Receive the local rainfall probability from multiple vehicle terminals within the area corresponding to the edge device.

[0082] In practice, the edge device will establish a communication connection with each vehicle within its signal coverage area. In this way, the vehicle will send its determined local rainfall probability (for example, the local rainfall probability is obtained through the process of steps 201 to 203 above) to the edge device in real time.

[0083] Then the edge device will collect the local rainfall probabilities sent by each vehicle and combine them.

[0084] Step 302: Based on the local rainfall probability sent from each vehicle terminal, determine the feature data corresponding to the area range.

[0085] In practice, the local rainfall probability sent by each vehicle at the current time t is used. This local rainfall probability also carries the location information of the corresponding sending vehicle. Based on the location information of each vehicle and the local rainfall probability, a graph can be constructed to obtain the regional weather map corresponding to the area. Based on the regional weather map, feature determination processing can be performed to determine one or more feature data corresponding to the area.

[0086] This feature data is a feature that can characterize the rainfall characteristics of the corresponding area.

[0087] Step 303: Input the feature data into the pre-trained regional rainfall recognition model to perform regional rainfall analysis and obtain regional rainfall recognition results.

[0088] In practice, the regional rainfall recognition model (e.g., an LSTM model, a bidirectional LSTM model, or a convolutional neural network model, preferably an LSTM model) is the regional rainfall recognition model received from the cloud server. The cloud server uses a second training sample set to train the pre-built initial regional rainfall recognition model, thereby obtaining a regional rainfall recognition model capable of performing regional rainfall analysis on feature data. Furthermore, the cloud server periodically or irregularly updates the regional rainfall recognition model and sends the updated model to the edge device, which then replaces the original regional rainfall recognition model with the updated one.

[0089] The regional rainfall identification model analyzes the feature data and outputs regional rainfall identification results, which represent the probability or degree of rainfall within the area corresponding to the edge device.

[0090] Step 304: Send the regional rainfall identification results to each vehicle terminal within the region.

[0091] In practice, after the edge device obtains the regional rainfall identification result, it sends the result to each vehicle within the area. Upon receiving the regional rainfall identification result, the corresponding vehicle corrects the local rainfall probability based on the result, resulting in a more accurate local rainfall probability. This, in turn, allows for safer vehicle control.

[0092] Through the above scheme, the edge device can integrate the local rainfall probabilities fed back by various vehicles within its assigned area and determine the feature data corresponding to that area. Since this feature data can characterize the rainfall characteristics of that area, the pre-trained regional rainfall recognition model can accurately analyze the regional rainfall using this feature data and output a regional rainfall recognition result that accurately indicates the rainfall probability or intensity within that area. After sending the regional rainfall recognition result to the corresponding vehicle, the vehicle can combine the local rainfall probability with the regional rainfall recognition result for more accurate driving control, thereby improving driving safety.

[0093] In some embodiments, step 302 includes: Step 3021: Based on the local rainfall probability sent from each vehicle terminal, determine the location rainfall probability corresponding to each location within the area.

[0094] In practice, the area will be divided into multiple locations, and each location may have no vehicle terminals or may have one or more vehicle terminals.

[0095] For locations without vehicle terminals, the corresponding location rainfall probability is set to empty; or, for locations with one vehicle terminal, the local rainfall probability of that vehicle terminal is weighted and used as the location rainfall probability; or, for locations with multiple vehicle terminals, the local rainfall probabilities of each vehicle terminal are weighted and summed to obtain the location rainfall probability.

[0096] Step 3022: Based on the location rainfall probability corresponding to each location, determine the feature data corresponding to the area range.

[0097] In practice, the probability of rainfall at each location within the region is graphically constructed to obtain a regional weather map. This allows for feature determination based on the regional weather map, resulting in feature data.

[0098] The above scheme can integrate and analyze the local rainfall probabilities fed back from various vehicles within the region, determine the accurate location rainfall probability for each location, and then construct a graph of the location rainfall probabilities corresponding to each location within the region, thereby obtaining a regional weather map that represents the rainfall situation in the region. Thus, the feature data determined based on the regional weather map is more accurate.

[0099] In some embodiments, step 3021 includes: Step 30211: Determine each location within the area and designate each location as a target location.

[0100] In a specific implementation, a two-dimensional coordinate system is constructed based on the region. The region is then divided into multiple locations on the two-dimensional coordinate system. Each location is traversed, and the traversed location is used as the target location to execute subsequent steps 30212 and 30213.

[0101] Step 30212: Determine at least one target vehicle terminal corresponding to the target location from among multiple vehicle terminals.

[0102] Step 30213: The local rainfall probabilities sent from at least one target vehicle are weighted and summed to obtain the location rainfall probability corresponding to the target location.

[0103] In practice, if no target vehicle terminal exists at the target location, the location rainfall probability corresponding to that target location will be set to a null value. For a target location with at least one target vehicle terminal, the local rainfall probability of that vehicle terminal is weighted and used as the location rainfall probability for a location with one vehicle terminal; or, for a location with multiple vehicle terminals, the local rainfall probabilities of each vehicle terminal corresponding to that location are weighted and summed to obtain the location rainfall probability.

[0104] The above scheme can traverse all locations within the region. For each target location traversed, the accurate location rainfall probability corresponding to that target location can be determined according to the above scheme. After the traversal is completed, the location rainfall probability corresponding to each location within the region can be obtained, which is convenient for subsequent feature filtering based on the location rainfall probability corresponding to each location.

[0105] In some embodiments, step 3022 includes: Step 30221: Divide the predetermined historical time range before the current moment into multiple historical time periods.

[0106] In practice, the edge device records the probability of rainfall at each location for a predetermined duration. After the predetermined duration ends, this predetermined duration is a predetermined historical time range before the current moment, which can be divided into multiple historical time periods according to fixed durations.

[0107] For example, the predetermined historical time range [t-15,t] before the current time t is divided into three historical time periods: [t-15,t-10), [t-10,t-5), and [t-5,t].

[0108] Step 30222: Based on the location rainfall probability corresponding to each location in each historical time period, determine the basic feature data corresponding to each historical time period.

[0109] In practice, this basic feature data represents the regional rainfall situation at various locations within the region.

[0110] The basic characteristic data includes at least one of the following: Regional Rainy Day Probability (P_region): The overall regional rainfall probability for a given historical time period. The regional rainy day probability (P_region) is obtained by weighted averaging of the location rainfall probabilities for each location within the corresponding historical time period.

[0111] Rate of change of probability: The slope of the linear regression of the probability of rainy days in the region during this historical period, reflecting whether the rainfall intensity is increasing (positive value indicates increasing) or decreasing (negative value indicates decreasing).

[0112] Data reliability is defined as the number of valid vehicle terminals participating in the calculation within the region during the historical time period (e.g., N_effective), or the weighted average confidence level of all vehicle terminal feedback within the region during the historical time period, or the weighted average of the data quality of all vehicle terminal feedback within the region during the historical time period. This data reliability characterizes the reliability of various data obtained within the region.

[0113] Spatial consistency: Within the region of this historical time period, the consistency of the local rainy day probability reported by each vehicle terminal is judged, and a consistency index is determined (e.g., the variance of the local rainy day probability reported by each vehicle terminal). The smaller the value of this spatial consistency, the more uniform the rainfall is in the region of this historical time period.

[0114] Step 30223: Combine the basic feature data of each historical time period to obtain the feature data corresponding to the area range.

[0115] In practice, the various historical time periods and their corresponding basic feature data are combined to construct a feature matrix, thereby obtaining accurate feature data corresponding to the regional scope. The dimension of this feature data is the number of historical time periods multiplied by the dimension of the basic feature data.

[0116] The above scheme combines basic feature data corresponding to multiple historical time periods, resulting in more comprehensive feature data that better reflects multi-dimensional rainfall characteristics, thus ensuring the accuracy and comprehensiveness of the feature data.

[0117] In some embodiments, after step 303, the method further includes determining a skid risk index for each vehicle end based on the regional rainfall identification results, the process of which is as follows: Step A1: Determine the regional rainfall level based on the regional rainfall identification results.

[0118] In practice, the rainfall identification result in the region can be either the regional rainfall level or the regional rainfall probability. The regional rainfall level is obtained by classifying the levels based on the regional rainfall probability.

[0119] Among them, the corresponding regional rainfall level (e.g., I rain The rainstorm warning system includes five levels: no rain, light rain, moderate rain, heavy rain, and torrential rain. The classification of these levels can be set according to actual needs; no specific limitations are made here.

[0120] Step A2: Designate each vehicle within the specified area as the target vehicle and execute the following: A21, obtain the road surface temperature and dew point temperature corresponding to the target vehicle end, and determine the temperature difference between the road surface temperature and the dew point temperature.

[0121] In practice, the road surface temperature (e.g., T) road This refers to the temperature data of the road surface where the target vehicle is located, detected by the temperature sensor on the target vehicle. Dew point temperature (e.g., T). dew,The dewpoint temperature (DT) is the temperature at which the air reaches saturation and condenses into liquid water, as detected by a dew point temperature sensor installed on the target vehicle. This temperature is determined by the air moisture content and pressure at the target vehicle's location.

[0122] Calculate the temperature difference between the road surface temperature and the dew point temperature. This temperature difference can characterize the rainfall situation at the location of the target vehicle.

[0123] A22, obtain the tire wear coefficient corresponding to the target vehicle end, and determine the degree of tire wear based on the tire wear coefficient.

[0124] In practice, the tire wear coefficient (e.g., W tire To characterize the wear condition of the tires on the vehicle corresponding to the target vehicle, the value range is [0,1]. The larger the tire wear coefficient, the lower the tire wear degree. For example, 0 is completely worn and 1 is brand new and has never been worn.

[0125] Then, subtracting the tire wear coefficient from 1 will give you the tire's unworn condition. For example, 1- W tire .

[0126] A23, the weighted sum of the regional rainfall level, the temperature difference, and the tire wear level is used to obtain the skid risk index of the target vehicle.

[0127] In practice, the weighting coefficients for regional rainfall level, temperature difference, and tire wear level are determined, and then a weighted sum is performed using these weighting coefficients to obtain the final skid risk index (e.g., R risk ).

[0128] R risk = α I rain + β (T) road -T dew )+ γ (1-W) tire ), among which, I rain For regional rainfall levels, for example, I rain ∈{1,2,3,4,5}, α This is a weighting coefficient for regional rainfall levels. β The weighting coefficient for the temperature difference. γ This is a weighting factor for the degree of tire wear. α + β + γ=1, the units of the parameters in the formula are not included in the calculation.

[0129] A24, the slip risk index is sent to the target vehicle.

[0130] In practice, after the edge device obtains the slip risk index according to the above process, it sends the slip risk index to the target vehicle, so that the target vehicle can perform driving control based on the slip risk index.

[0131] In addition, the edge device can directly send the determined rainfall level of the area to the target vehicle, and the target vehicle can execute the above steps A21 to A23, and then the target vehicle can perform driving control based on the slip risk index.

[0132] The above scheme combines the regional rainfall level, the temperature difference between the road surface temperature and dew point temperature at the target vehicle, and the tire wear coefficient at the target vehicle to determine an accurate skid risk index for the vehicle. After the skid risk index is sent to the target vehicle, the target vehicle can perform accurate driving control based on the skid risk index, thereby improving the safety of the target vehicle in controlling the vehicle's driving.

[0133] In some embodiments, after step A23, the method further includes: Step B1: Obtain the reference road surface adhesion coefficient at the location of the target vehicle end, and determine the current road surface adhesion coefficient based on the reference road surface adhesion coefficient and the slip risk index.

[0134] In practice, the edge device can retrieve the reference road surface adhesion coefficient corresponding to the location of the target vehicle. This reference road surface adhesion coefficient is the adhesion coefficient corresponding to the dry road surface at that location (for example, set to 0.8). In this way, the current road surface adhesion coefficient can be calculated based on the slip risk index obtained above.

[0135] The formula for the current road surface adhesion coefficient μ is μ = μ0 (1 - R risk ), where μ0 is the reference road surface adhesion coefficient.

[0136] Step B2: Obtain the current speed of the target vehicle and determine the safe following distance based on the current speed and the current road surface adhesion coefficient.

[0137] In practice, the safe vehicle distance D safe The calculation formula is: Where v is the current vehicle speed, The driver's reaction time is given (e.g., set to 1.5 s), and g is the gravitational acceleration (e.g., 9.8 m / s²).

[0138] Step B3: Send the safe distance to the target vehicle so that the target vehicle can perform driving control based on the safe distance.

[0139] In practice, once the edge device obtains the safe distance, it can send the safe distance to the target vehicle. This allows the target vehicle to control its driving according to the safe distance, ensuring that the distance between the vehicle and the vehicle in front is always greater than the safe distance, thus guaranteeing safe driving.

[0140] Alternatively, the process of steps B1 to B2 above can also be executed by the target vehicle. In this way, the target vehicle (i.e., the vehicle execution entity of the embodiment executed by steps 201 to 203 above) can control the distance between the vehicle and the vehicle in front to always be greater than the safe distance according to the safe distance, so as to ensure the safe driving of the vehicle.

[0141] The above scheme ensures that the safe following distance can be accurately calculated based on the slip risk index corresponding to the target vehicle and the reference road surface adhesion coefficient at the location of the target vehicle. After the safe following distance is sent to the target vehicle, the target vehicle can control the distance between itself and the vehicle in front to always be greater than the safe following distance, thus ensuring the safe driving of the vehicle.

[0142] In a preferred embodiment, the edge device also connects to a cloud server, which continuously updates and trains the regional rainfall recognition model. The specific update and training process is as follows: (1) Collect sample feature data corresponding to each area (the process of obtaining the sample feature data is the same as the process of obtaining the feature data mentioned above), and mark the sample feature data with the corresponding real regional rainfall identification results to obtain the second training sample. Combine the second training samples to form the second training sample set.

[0143] (2) Input the second training samples in the second training sample set into the regional rainfall recognition model (e.g., LSTM model). The rainfall probability determination model has a regional rainy day probability analysis layer, which is used to analyze the regional rainy day probability of each historical time period in the second training sample to determine the first regional rainfall probability.

[0144] (3) The rainfall identification model in this region is equipped with a probability change rate analysis layer, which is used to analyze the probability change rate of each historical time period in the second training sample to determine the rainfall probability in the second region.

[0145] (4) The rainfall identification model in this region is equipped with a data reliability analysis layer, which is used to analyze the data reliability of each historical time period in the second training sample to determine the rainfall probability of the third region.

[0146] (5) The rainfall identification model in this region is equipped with a spatial consistency analysis layer, which is used to analyze the spatial consistency of each historical time period in the second training sample to determine the rainfall probability of the fourth region.

[0147] (6) The rainfall identification model in this region is equipped with a fusion analysis layer, which is used to weight and sum the rainfall probabilities of the first region, the second region, the third region, and the fourth region to obtain the rainfall identification result of the training region.

[0148] (7) Compare the rainfall recognition results of the training area with the real rainfall recognition results of the marked area, determine the second loss function based on the comparison results, calculate the corresponding second loss value, update and optimize the parameters of each layer of the regional rainfall recognition model based on the second loss value, and obtain the optimized regional rainfall recognition model.

[0149] (8) For each second training sample in the second training sample set, repeat the process from (2) to (7) based on the previously optimized regional rainfall recognition model until the second training sample set is trained, or the final optimized regional rainfall recognition model reaches the predetermined convergence level, or the accuracy of the final optimized regional rainfall recognition model reaches the predetermined accuracy threshold, then stop training and use the final optimized regional rainfall recognition model as the final optimized regional rainfall recognition model.

[0150] (9) Send the final optimized regional rainfall identification model to the edge device so that the edge device can replace the original regional rainfall identification model with the final optimized regional rainfall identification model for rainfall analysis.

[0151] It should be noted that the method in this embodiment can be executed by a single device, such as a computer or server. The method can also be applied in a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method in this embodiment, and the multiple devices will interact with each other to complete the method described.

[0152] It should be noted that the above description describes some embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0153] Based on the same inventive concept, corresponding to the weather recognition method applied to the vehicle end in any of the above embodiments, this application also provides a weather recognition device installed on the vehicle end.

[0154] refer to Figure 4 The weather recognition device includes: The environmental data acquisition module 401 is configured to acquire environmental image data, environmental audio data, and environmental perception data around the vehicle. The local rainfall analysis module 402 is configured to input the environmental image data, the environmental audio data and the environmental perception data into a predetermined rainfall probability determination model to perform rainfall analysis and obtain the local rainfall probability of the vehicle. The local rainfall probability sending module 403 is configured to send the local rainfall probability to an edge device so that the edge device can perform regional rainfall identification based on the local rainfall probability.

[0155] In some embodiments, the local rainfall analysis module 402 is specifically configured as follows: After inputting the environmental image data, the environmental audio data, and the environmental perception data into a pre-determined rainfall probability determination model, the execution process of the rainfall probability determination model is as follows: Rainfall analysis is performed based on the environmental image data to determine a first probability of rainfall. Based on the environmental audio data, rainfall analysis is performed to determine the second rainfall probability; Based on the environmental perception data, rainfall analysis is performed to determine the third rainfall probability; The local rainfall probability of the vehicle is obtained by weighted summing of the first rainfall probability, the second rainfall probability, and the third rainfall probability.

[0156] In some embodiments, the local rainfall analysis module 402 is further configured as follows: Obtain the image confidence, audio confidence, and environmental perception confidence of the previous moment output by the rainfall probability determination model at the previous moment; Based on the image confidence score of the previous moment, determine the image weight value of the environmental image data; Based on the audio confidence level of the previous moment, determine the audio weight value of the environmental audio data; Based on the environmental perception confidence level of the previous moment, determine the perception weight value of the environmental perception data; The image weight value is multiplied by the first rainfall probability to obtain a first product, the audio weight value is multiplied by the second rainfall probability to obtain a second product, the perception weight value is multiplied by the third rainfall probability to obtain a third product, and the first product, the second product and the third product are added together to obtain the local rainfall probability of the vehicle. The image confidence level at the current moment is determined based on the environmental image data, the audio confidence level at the current moment is determined based on the environmental audio data, and the environmental perception confidence level at the current moment is determined based on the environmental perception data. The rainfall probability determination model outputs the current moment's image confidence, the current moment's audio confidence, and the current moment's environmental perception confidence, along with the local rainfall probability.

[0157] Based on the same inventive concept, corresponding to the weather recognition method applied to edge devices in any of the above embodiments, this application also provides a weather recognition device disposed on an edge device.

[0158] refer to Figure 5 The weather recognition device includes: The receiving module 501 is configured to receive local rainfall probabilities from multiple vehicle terminals within the area corresponding to the edge device; The feature data determination module 502 is configured to determine the feature data corresponding to the area range based on the local rainfall probability sent from each vehicle terminal. The regional rainfall identification module 503 is configured to input the feature data into a pre-trained regional rainfall identification model to perform regional rainfall analysis and obtain regional rainfall identification results; The regional rainfall identification result sending module 504 is configured to send the regional rainfall identification result to each vehicle terminal within the regional area.

[0159] In some embodiments, the feature data determination module 502 is specifically configured to: Based on the local rainfall probability sent from each vehicle terminal, the location rainfall probability corresponding to each location within the area is determined. Based on the location-specific rainfall probability at each location, characteristic data corresponding to the area range are determined.

[0160] In some embodiments, the feature data determination module 502 is further configured to: Determine each location within the defined area and designate each location as a target location. From multiple vehicle terminals, determine at least one target vehicle terminal corresponding to the target location; The local rainfall probabilities sent from at least one target vehicle are weighted and summed to obtain the location rainfall probability corresponding to the target location.

[0161] In some embodiments, the feature data determination module 502 is further configured to: Divide the predetermined historical time range before the current moment into multiple historical time periods; Based on the probability of rainfall at each location in each historical time period, the basic feature data corresponding to each historical time period are determined. The basic feature data of each historical time period are combined to obtain the feature data corresponding to the region.

[0162] In some embodiments, the device further includes a slip risk identification module, configured to: After obtaining the regional rainfall identification results, the regional rainfall level is determined based on the regional rainfall identification results; Treat each vehicle within the defined area as a target vehicle and execute the following: Obtain the road surface temperature and dew point temperature corresponding to the target vehicle end, and determine the temperature difference between the road surface temperature and the dew point temperature; Obtain the tire wear coefficient corresponding to the target vehicle end, and determine the degree of tire wear based on the tire wear coefficient; The slip risk index of the target vehicle is obtained by weighted summing of the rainfall level in the area, the temperature difference, and the tire wear level. The slip risk index is sent to the target vehicle.

[0163] In some embodiments, the apparatus further includes a safe distance determination module, configured to: After obtaining the skid risk index of the target vehicle end, the reference road surface adhesion coefficient at the location of the target vehicle end is obtained, and the current road surface adhesion coefficient is determined based on the reference road surface adhesion coefficient and the skid risk index. Obtain the current speed of the target vehicle and determine a safe following distance based on the current speed and the current road surface adhesion coefficient; The safe distance is sent to the target vehicle so that the target vehicle can perform driving control based on the safe distance.

[0164] For ease of description, the above devices are described in terms of function, divided into various modules. Of course, in implementing this application, the functions of each module can be implemented in one or more software and / or hardware.

[0165] The apparatus of the above embodiments is used to implement the corresponding method in any of the foregoing embodiments and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0166] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the methods described in any of the above embodiments.

[0167] Figure 6 This embodiment illustrates a more specific hardware structure of an electronic device. The device may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.

[0168] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0169] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0170] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0171] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0172] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0173] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0174] The electronic devices described above are used to implement the corresponding methods in any of the foregoing embodiments and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0175] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides a non-transitory computer-readable storage medium that stores computer instructions for causing the computer to perform the methods described in any of the above embodiments.

[0176] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random-access memory (SRAM), dynamic random-access memory (DRAM), other types of random-access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital video disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0177] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to perform the methods described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0178] Based on the same concept, corresponding to any of the above embodiments, this application also provides a computer program product, including computer program instructions, which, when run on a computer, cause the computer to perform the method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0179] Based on the same inventive concept, this application also provides a system including the vehicle terminal, and / or edge device, and / or cloud server in the above embodiments. Specifically, the weather recognition method executed by the vehicle terminal, the weather recognition method executed by the edge device, and the model update process executed by the cloud server, and the corresponding benefits are not described here.

[0180] Based on the same inventive concept, this application also provides a vehicle, including the vehicle-mounted terminal or the electronic device described in the above embodiments. The beneficial effects of embodiments having corresponding vehicle-mounted terminals or electronic devices will not be elaborated further here.

[0181] It is understood that before using the technical solutions of the various embodiments in this application, users will be informed of the type, scope of use, and usage scenarios of the personal information involved in an appropriate manner, and user authorization will be obtained.

[0182] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose, based on the prompt message, whether to provide personal information to the software or hardware such as electronic devices, applications, servers, or storage media performing the operations described in this application.

[0183] As an optional but not limited implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0184] It is understood that the above notification and user authorization process is merely illustrative and does not limit the implementation of this application. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this application.

[0185] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this application (including the claims) is limited to these examples; within the framework of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this application as described above, which are not provided in the details for the sake of brevity.

[0186] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this application, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this application, and this also takes into account the fact that the details of the implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this application will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of this application, it will be apparent to those skilled in the art that the embodiments of this application can be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0187] Although this application has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0188] The embodiments of this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this application should be included within the protection scope of this application.

Claims

1. A weather identification method, characterized in that, Applied to the vehicle end, the method includes: Acquire environmental image data, environmental audio data, and environmental perception data around the vehicle; The environmental image data, the environmental audio data, and the environmental perception data are input into a pre-determined rainfall probability determination model to perform rainfall analysis and obtain the local rainfall probability for the vehicle. The local rainfall probability is sent to the edge device so that the edge device can identify regional rainfall based on the local rainfall probability.

2. The method according to claim 1, characterized in that, The step of inputting the environmental image data, the environmental audio data, and the environmental perception data into a pre-determined rainfall probability determination model to perform rainfall analysis and obtain the local rainfall probability for the vehicle includes: After inputting the environmental image data, the environmental audio data, and the environmental perception data into a pre-determined rainfall probability determination model, the execution process of the rainfall probability determination model is as follows: Rainfall analysis is performed based on the environmental image data to determine a first probability of rainfall. Based on the environmental audio data, rainfall analysis is performed to determine the second rainfall probability; Based on the environmental perception data, rainfall analysis is performed to determine the third rainfall probability; The local rainfall probability of the vehicle is obtained by weighted summing of the first rainfall probability, the second rainfall probability, and the third rainfall probability.

3. The method according to claim 2, characterized in that, The step of weighted summing of the first rainfall probability, the second rainfall probability, and the third rainfall probability to obtain the local rainfall probability for the vehicle includes: Obtain the image confidence, audio confidence, and environmental perception confidence of the previous moment output by the rainfall probability determination model at the previous moment; Based on the image confidence score of the previous moment, determine the image weight value of the environmental image data; Based on the audio confidence level of the previous moment, determine the audio weight value of the environmental audio data; Based on the environmental perception confidence level of the previous moment, determine the perception weight value of the environmental perception data; The image weight value is multiplied by the first rainfall probability to obtain a first product, the audio weight value is multiplied by the second rainfall probability to obtain a second product, the perception weight value is multiplied by the third rainfall probability to obtain a third product, and the first product, the second product and the third product are added together to obtain the local rainfall probability of the vehicle. The image confidence level at the current moment is determined based on the environmental image data, the audio confidence level at the current moment is determined based on the environmental audio data, and the environmental perception confidence level at the current moment is determined based on the environmental perception data. The rainfall probability determination model outputs the current moment's image confidence, the current moment's audio confidence, and the current moment's environmental perception confidence, along with the local rainfall probability.

4. A weather identification method, characterized in that, Applied to edge devices, the method includes: Receive local rainfall probability data from multiple vehicle terminals within the area corresponding to the edge device; Based on the local rainfall probability sent from each vehicle terminal, the characteristic data corresponding to the area range is determined; The feature data is input into a pre-trained regional rainfall recognition model to perform regional rainfall analysis and obtain regional rainfall recognition results. The rainfall identification results for the area are sent to each vehicle within the area.

5. The method according to claim 4, characterized in that, The determination of feature data corresponding to the area range based on the local rainfall probability sent from each vehicle terminal includes: Based on the local rainfall probability sent from each vehicle terminal, the location rainfall probability corresponding to each location within the area is determined. Based on the location-specific rainfall probability at each location, characteristic data corresponding to the area range are determined.

6. The method according to claim 5, characterized in that, The determination of the location rainfall probability for each location within the area based on the local rainfall probability sent from each vehicle terminal includes: Determine each location within the defined area and designate each location as a target location. From multiple vehicle terminals, determine at least one target vehicle terminal corresponding to the target location; The local rainfall probabilities sent from at least one target vehicle are weighted and summed to obtain the location rainfall probability corresponding to the target location.

7. The method according to claim 5, characterized in that, The determination of feature data corresponding to the area range based on the location rainfall probability at each location includes: Divide the predetermined historical time range before the current moment into multiple historical time periods; Based on the probability of rainfall at each location in each historical time period, the basic feature data corresponding to each historical time period are determined. The basic feature data of each historical time period are combined to obtain the feature data corresponding to the region.

8. The method according to claim 4, characterized in that, After obtaining the regional rainfall identification results, the following is also included: Based on the regional rainfall identification results, the regional rainfall level is determined; Treat each vehicle within the defined area as a target vehicle and execute the following: Obtain the road surface temperature and dew point temperature corresponding to the target vehicle end, and determine the temperature difference between the road surface temperature and the dew point temperature; Obtain the tire wear coefficient corresponding to the target vehicle end, and determine the degree of tire wear based on the tire wear coefficient; The slip risk index of the target vehicle is obtained by weighted summing of the rainfall level in the area, the temperature difference, and the tire wear level. The slip risk index is sent to the target vehicle.

9. The method according to claim 8, characterized in that, After obtaining the skid risk index of the target vehicle end, the method further includes: Obtain the reference road surface adhesion coefficient at the location of the target vehicle end, and determine the current road surface adhesion coefficient based on the reference road surface adhesion coefficient and the slip risk index; Obtain the current speed of the target vehicle and determine a safe following distance based on the current speed and the current road surface adhesion coefficient; The safe distance is sent to the target vehicle so that the target vehicle can perform driving control based on the safe distance.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 9.

Images