Traffic accident early warning method, device and equipment in bad weather and readable medium

Abstract

Embodiments of the present disclosure disclose a traffic accident warning method and device in adverse weather, equipment and readable medium. A specific embodiment of the method comprises: obtaining a multi-modal data set corresponding to a target driver; performing fusion processing on the multi-modal data set to generate a fusion data matrix; performing driver feature extraction processing on the fusion data matrix to generate an extracted feature set; based on the extracted feature set, evaluating the behavior ability of the target driver to generate a driver behavior evaluation value; selecting a traffic accident warning model corresponding to the extracted feature set from a set of traffic accident warning models as a target warning model, and inputting the fusion data matrix into the target warning model to obtain an accident warning result; based on the driver behavior evaluation value and the accident warning result, generating an accident warning report, and sending the accident warning report to a target terminal for warning operation. The embodiment improves the driving safety of the driver.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the field of computer technology, and more specifically to a method, apparatus, device, and readable medium for traffic accident early warning under severe weather conditions. Background Technology

[0002] Traffic accident prevention technologies can provide traffic warnings before a driver is involved in an accident, thereby ensuring the safety of drivers and roads. Under the influence of severe weather (such as typhoons or snowstorms), drivers are highly susceptible to traffic accidents, resulting in serious consequences such as vehicle damage or even driver injury. Currently, the common approach to traffic accident prevention is to analyze driving data (e.g., historical traffic accident data or driver violation data) and issue warnings.

[0003] However, when using the above methods for traffic accident prevention, the following technical problems often arise:

[0004] Data analysis based solely on single-dimensional data is not comprehensive enough. It fails to consider the impact of severe weather on drivers and cannot establish a dynamic matching mechanism between driver profiles and regional risk characteristics. This leads to biased warning results and lower driving safety for drivers.

[0005] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion that follows. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0007] Some embodiments of this disclosure provide methods, apparatus, electronic devices, and computer-readable media for traffic accident early warning in severe weather to address one or more of the technical problems mentioned in the background section above.

[0008] In a first aspect, some embodiments of this disclosure provide a method for traffic accident early warning under severe weather conditions. The method includes: in response to receiving driving monitoring information for a target driver sent by a target terminal, acquiring a multimodal dataset corresponding to the target driver; performing fusion processing on the multimodal dataset to generate a fusion data matrix; performing driver feature extraction processing on the fusion data matrix to generate an extracted feature set; evaluating the target driver's behavioral capabilities based on the extracted feature set to generate a driver behavior evaluation value; selecting a traffic accident early warning model corresponding to the extracted feature set from a set of traffic accident early warning models as a target early warning model; inputting the fusion data matrix into the target early warning model to obtain an accident early warning result; generating an accident early warning report based on the driver behavior evaluation value and the accident early warning result; and sending the accident early warning report to the target terminal for early warning operation.

[0009] Secondly, some embodiments of this disclosure provide a traffic accident early warning device under severe weather conditions. The device includes: an acquisition unit configured to acquire a multimodal dataset corresponding to the target driver in response to receiving driving monitoring information for the target driver sent by a target terminal; a fusion unit configured to perform fusion processing on the multimodal dataset to generate a fused data matrix; a feature extraction unit configured to perform driver feature extraction processing on the fused data matrix to generate an extracted feature set; an evaluation unit configured to evaluate the behavioral ability of the target driver based on the extracted feature set to generate a driver behavior evaluation value; a selection unit configured to select a traffic accident early warning model corresponding to the extracted feature set from a traffic accident early warning model set as a target early warning model, and input the fused data matrix into the target early warning model to obtain an accident early warning result; and a generation unit configured to generate an accident early warning report based on the driver behavior evaluation value and the accident early warning result, and send the accident early warning report to the target terminal for early warning operation.

[0010] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0011] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0012] The various embodiments of this disclosure have the following beneficial effects: the traffic accident early warning method under severe weather conditions according to some embodiments of this disclosure improves the driving safety of drivers. Specifically, the reason for the low driving safety of drivers is that when data analysis is performed using only single-dimensional data, the data content is not comprehensive enough, the impact of severe weather on driving is not considered, and a dynamic matching mechanism between driver profiles and regional risk characteristics cannot be established, resulting in biased early warning results and low driving safety of drivers. Based on this, the traffic accident early warning method under severe weather conditions according to some embodiments of this disclosure firstly, in response to receiving driving monitoring information for a target driver sent by a target terminal, obtains a multimodal dataset corresponding to the target driver. Thus, multi-dimensional data of the driver can be obtained. Secondly, the multimodal dataset is fused to generate a fused data matrix. Thus, data from different dimensions can be unified to the same dimension through data fusion. Then, driver feature extraction processing is performed on the fused data matrix to generate an extracted feature set. Thus, various data features of the multimodal data can be extracted. Afterwards, based on the extracted feature set, the behavioral ability of the target driver is evaluated to generate a driver behavior evaluation value. Thus, a driver ability evaluation can be generated to determine the driver's driving ability. Next, a traffic accident early warning model corresponding to the extracted feature set is selected from the traffic accident early warning model set as the target early warning model. The fused data matrix is ​​then input into the target early warning model to obtain the accident early warning result. This allows for the prediction of road traffic accidents. Finally, based on the driver behavior evaluation value and the accident early warning result, an accident early warning report is generated and sent to the target terminal for early warning operation. This completes the early warning of traffic accidents. Furthermore, by fusing and analyzing multimodal driver data, a dynamic matching mechanism between driver profiles and regional risk characteristics can be established, generating accurate early warning results and improving driver safety. Attached Figure Description

[0013] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0014] Figure 1 This is a flowchart of some embodiments of the traffic accident early warning method under severe weather conditions according to the present disclosure;

[0015] Figure 2 This is a schematic diagram of the structure of some embodiments of the traffic accident warning device under severe weather conditions according to the present disclosure;

[0016] Figure 3 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0017] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0018] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0019] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0020] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0021] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0022] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] Figure 1 A flow 100 of some embodiments of a traffic accident warning method under severe weather conditions according to the present disclosure is shown. The traffic accident warning method under severe weather conditions includes the following steps:

[0024] Step 101: In response to receiving driving monitoring information for the target driver sent by the target terminal, obtain the corresponding multimodal dataset of the target driver.

[0025] In some embodiments, the implementing entity (e.g., a server) of the traffic accident early warning method under severe weather conditions can obtain a multimodal dataset corresponding to the target driver in response to receiving driving monitoring information for the target driver sent by the target terminal. In practice, the implementing entity can obtain the multimodal dataset of the target driver through an associated terminal interface. The associated terminal interface can be an interface for connecting to various associated servers. The multimodal data may include, but is not limited to: driver training data, vehicle violation data, vehicle accident data, and areas with frequent violations and accidents. The driving monitoring information can represent driving monitoring of a specific driver who is currently driving.

[0026] Optionally, before step 102, the following steps are also included:

[0027] The first step is to clean each multimodal data in the above multimodal dataset to generate a cleaned dataset.

[0028] In some embodiments, the execution entity may perform cleaning processing on each multimodal data in the multimodal dataset to generate a cleaned dataset. This cleaning processing may involve removing outlier data from the multimodal dataset.

[0029] The second step is to perform feature alignment on the cleaned dataset to generate an aligned dataset, which serves as a multimodal dataset.

[0030] In some embodiments, the execution entity may perform feature alignment processing on the cleaned dataset to generate an aligned dataset as a multimodal dataset.

[0031] Step 102: Perform fusion processing on the multimodal dataset to generate a fused data matrix.

[0032] In some embodiments, the aforementioned execution entity may perform fusion processing on the aforementioned multimodal dataset to generate a fused data matrix.

[0033] In practice, multimodal datasets can be fused using the following steps to generate a fused data matrix:

[0034] The first step is to determine the confidence level of each multimodal data point in the multimodal dataset to generate a data confidence level.

[0035] The second step involves fusing the multimodal data in the aforementioned multimodal dataset based on the generated confidence scores of each data point, in order to generate a fused data matrix.

[0036] Step 103: Perform driver feature extraction processing on the fused data matrix to generate an extracted feature set.

[0037] In some embodiments, the execution entity may perform driver feature extraction processing on the fused data matrix to generate an extracted feature set.

[0038] Step 104: Based on the extracted feature set, evaluate the target driver's behavioral ability to generate a driver behavior evaluation value.

[0039] In some embodiments, the executing entity may evaluate the behavioral capabilities of the target driver based on the extracted feature set to generate a driver behavior evaluation value.

[0040] In some optional implementations of certain embodiments, the aforementioned executing entity may perform a behavioral ability evaluation on the target driver through the following steps to generate a driver behavior evaluation value:

[0041] The first step is to sort the extracted features in the above feature set according to their timeliness to generate an extracted feature sequence. In practice, the timeliness can be sorted according to the acquisition time of the multimodal data corresponding to the extracted features.

[0042] The second step is to determine the weight feature value corresponding to each extracted feature in the above extracted feature sequence to generate a weighted feature sequence. Here, the timeliness weight corresponding to each extracted feature can be determined based on a pre-set timeliness weight rule.

[0043] The third step is to generate a driver behavior capacity table.

[0044] Fourth, for each driver's behavioral ability in the above driver behavioral ability table, perform the following processing steps:

[0045] The first processing step is to determine at least one extractable feature corresponding to the aforementioned driver's behavioral capabilities. These extractable features may include, but are not limited to: driver's age, driver's medical history, driver's physiological data, and driving experience.

[0046] The second processing step involves predicting the driver's behavioral capabilities based on at least one of the extracted features to generate a capability prediction value.

[0047] The third processing step is to determine the scene type corresponding to the aforementioned multimodal dataset. This scene type can be the driving scenario of the vehicle driven by the target driver. For example, the scene type may include, but is not limited to: heavy rain, nighttime driving, highways, and school zones.

[0048] The fourth processing step involves correcting the predicted capabilities based on the aforementioned scenario types to generate corrected capability values. In practice, a correction factor corresponding to the scenario type can be determined, and the product of the predicted capabilities and the correction factor can be used to determine the corrected capability value.

[0049] The fifth step involves generating an intervention strategy matrix corresponding to the target driver based on the generated corrected ability values, and determining the sum of the generated corrected ability values ​​as the driver behavior evaluation value. In practice, for each of the above corrected ability values, a preset strategy corresponding to that corrected ability value can be selected from a preset strategy set. The selected preset strategies are then combined with the corrected ability values ​​to form an intervention strategy matrix.

[0050] The relevant content in steps one through five above serves as an inventive point of this disclosure. Combined with step "106" below, it solves the technical problem that "when determining a driver's behavioral ability, the driver's driving history information (e.g., driving experience, medical history, etc.) is typically used. However, because driving history information changes over time, the predicted behavioral evaluation based on this information can be biased, thus affecting the prediction results of traffic accidents and resulting in lower safety for vehicles and drivers." Factors contributing to lower safety for vehicles and drivers often include: when determining a driver's behavioral ability, the driver's driving history information (e.g., driving experience, medical history, etc.) is typically used. However, because driving history information changes over time, the predicted behavioral evaluation based on this information can be biased, thus affecting the prediction results of traffic accidents and resulting in lower safety for vehicles and drivers. Solving these factors can improve the safety of vehicles and drivers. To achieve this, firstly, the timeliness of each extracted feature in the aforementioned feature set is ranked to generate an extracted feature sequence. Therefore, the timeliness of the extracted features can be ranked according to their time, identifying the most timely extracted features. Second, determine the weighted feature value corresponding to each extracted feature in the above extracted feature sequence to generate a weighted feature sequence. This allows the determination of the weight of each extracted feature. Third, generate a driver behavior ability table; for each driver behavior ability in the above driver behavior ability table, perform the following processing steps: First, determine at least one extracted feature corresponding to the above driver behavior ability. This allows the determination of the extracted feature corresponding to each driver behavior ability. Second, based on the above at least one extracted feature, perform ability prediction on the above driver behavior ability to generate an ability prediction value. This allows the generation of a corresponding ability prediction for each driver behavior ability. Then, determine the scene type corresponding to the above multimodal dataset; based on the above scene type, perform correction processing on the above ability prediction value to generate a corrected ability value. This allows the driver's behavior ability prediction to be corrected based on the current scene type. Finally, based on the generated corrected ability values, generate an intervention strategy matrix corresponding to the above target driver, and determine the sum of the generated corrected ability values ​​as the driver behavior evaluation value. This allows the comprehensive generation of the driver's behavior evaluation value. In conjunction with "Step 106," based on the aforementioned driver behavior evaluation value and accident warning results, an accident warning report is generated and sent to the aforementioned target terminal for warning operation. This allows for warning operations to be initiated for drivers with low behavior evaluation values, thereby preventing traffic accidents and improving the safety of vehicles and drivers.

[0051] Step 105: Select the traffic accident early warning model corresponding to the extracted feature set from the traffic accident early warning model set as the target early warning model, and input the fused data matrix into the target early warning model to obtain the accident early warning result.

[0052] In some embodiments, the aforementioned execution entity can select a traffic accident early warning model corresponding to the extracted feature set from the traffic accident early warning model set as the target early warning model, and input the aforementioned fused data matrix into the target early warning model to obtain the accident early warning result. The aforementioned traffic accident early warning model can be a pre-trained long short-term memory artificial neural network model. Here, the aforementioned traffic accident early warning model can include a five-layer structure. The first layer, the input layer, is used to receive the fused data matrix. The second layer, the hidden layer, includes three hidden layers, used to extract the temporal features of driving behavior from the fused data matrix. The third layer, the attention mechanism layer, is used to weight the key features in the temporal features of driving behavior. The aforementioned temporal features of driving behavior can include, but are not limited to: sharp turns, rapid acceleration and deceleration. The fourth layer, the fully connected layer, is used to map the weighted features to the accident probability space. The fifth layer, the output layer, outputs the accident occurrence probability value through an activation function.

[0053] In some optional implementations of certain embodiments, the aforementioned execution entity may determine the target early warning model through the following steps:

[0054] The first step is to select at least one feature representing the environment from the above-mentioned feature set as the target extraction feature, thus obtaining the target extraction feature set. The target extraction features in the above-mentioned target extraction feature set may include, but are not limited to: visibility, vehicle distance, and road surface condition.

[0055] The second step involves extracting a feature set based on the aforementioned target, and selecting preset scene feature information that meets preset similarity conditions from a preset scene feature information set as the target scene feature information. The preset scene feature information in the aforementioned preset scene feature information set corresponds to a traffic accident warning model in the aforementioned traffic accident warning model set. The preset scene feature information in the aforementioned preset scene feature information set can be features of a traffic scene corresponding to a specific traffic accident warning model.

[0056] The third step is to determine the scene similarity between the aforementioned target extracted feature set and the aforementioned target scene feature information. In practice, the scene similarity between each target extracted feature in the target extracted feature set and the target scene feature information can be determined using Euclidean distance.

[0057] The fourth step involves acquiring the fused data matrix in real time when the scene similarity is less than or equal to a preset similarity threshold. The preset similarity threshold can be a pre-defined threshold for scene similarity.

[0058] Fifth, extract at least one temporal feature from the fused data matrix to generate a temporal feature set. The temporal features in this set may include, but are not limited to, vehicle speed and steering wheel angle within a historical time period. The historical time period can be the time segment representing the period from a point before the current time to the current time.

[0059] Step 6: Based on the aforementioned temporal feature set, incrementally train the basic scene model to generate a sudden scene model that extracts features for the aforementioned targets. The aforementioned basic scene model can be an LSTM model.

[0060] Step 7: Based on the historical fusion data matrix set, determine the model accuracy corresponding to the above-mentioned emergency scenario model. In practice, the above-mentioned historical fusion data matrix set can be used as the test input set, and the early warning results corresponding to each historical fusion data matrix can be used as the test output set to determine the model accuracy corresponding to the above-mentioned emergency scenario model.

[0061] Step 8: In response to the above model accuracy being greater than or equal to the preset model accuracy, the above emergency scenario model is determined as the target early warning model, and the above emergency scenario model is added to the above traffic accident early warning model set.

[0062] Step 9: In response to the fact that the accuracy of the above model is less than the accuracy of the above preset model, the above emergency scenario model is determined as the target early warning model.

[0063] Step 10: Input the above fused data matrix into the above target early warning model to obtain the accident early warning result.

[0064] Step 11: Based on the above accident warning results, control the associated unmanned vehicles to intercept the vehicle driven by the target driver.

[0065] The relevant content in steps one through eleven above constitutes an inventive point of this disclosure, solving the technical problem that "when selecting a traffic warning model, because the preset traffic warning model cannot cover all traffic scenarios, the selected traffic warning model cannot fully provide safety warnings for vehicles, resulting in lower safety for vehicles and drivers." Factors contributing to lower safety for vehicles and drivers often include: when selecting a traffic warning model, because the preset traffic warning model cannot cover all traffic scenarios, the selected traffic warning model cannot fully provide safety warnings for vehicles, resulting in lower safety for vehicles and drivers. Solving these factors can improve the safety of vehicles and drivers. To achieve this, firstly, at least one extracted feature representing the environment is selected from the extracted feature set as a target extracted feature, resulting in a target extracted feature set. This allows the determination of the scene characteristics of the current vehicle's location. Secondly, based on the target extracted feature set, preset scene feature information that meets preset similarity conditions is selected from a preset scene feature information set as target scene feature information; the scene similarity between the target extracted feature set and the target scene feature information is determined. Therefore, target scene feature information similar to the vehicle's location can be identified, and the similarity can be determined. Third, in response to the scenario similarity being less than or equal to a preset similarity threshold, a fusion data matrix is ​​acquired in real time; at least one temporal feature is extracted from the fusion data matrix to generate a temporal feature set. This allows the acquisition of the vehicle's temporal features. Fourth, based on the temporal feature set, a basic scenario model is trained to generate a sudden scenario model with extracted features for the target. This generates a warning model for the current scenario. Fifth, based on the historical fusion data matrix set, the model accuracy corresponding to the sudden scenario model is determined; in response to the model accuracy being greater than or equal to a preset model accuracy, the sudden scenario model is identified as the target warning model and added to the traffic accident warning model set. This allows the trained model to be stored and maintained when the model accuracy is greater than or equal to the threshold. Sixth, in response to the model accuracy being less than the preset model accuracy, the sudden scenario model is identified as the target warning model. This allows the model to be used only as a temporary model for a single use when the model accuracy is less than the threshold. Seventh, the aforementioned fused data matrix is ​​input into the aforementioned target early warning model to obtain accident early warning results; based on the aforementioned accident early warning results, the associated unmanned vehicles are controlled to intercept the vehicle driven by the aforementioned target driver. Thus, by dynamically generating the model, the limitations of preset traffic early warning models in covering all traffic scenarios can be avoided, thereby intercepting vehicles posing driving hazards and improving the safety of vehicles and drivers.

[0066] Step 106: Based on the driver behavior evaluation value and the accident warning result, generate an accident warning report and send the accident warning report to the target terminal for warning operation.

[0067] In some embodiments, the aforementioned executing entity may generate an accident warning report based on the aforementioned driver behavior evaluation value and the aforementioned accident warning result, and send the aforementioned accident warning report to the aforementioned target terminal for warning operation.

[0068] In practice, the following steps can be used to generate an accident warning report based on driver behavior evaluation values ​​and accident warning results, and then send the accident warning report to the target terminal for warning operation:

[0069] The first step is to obtain an accident report template.

[0070] The second step is to fill in the above-mentioned correction capability values, intervention strategy matrix, driver behavior evaluation values, and accident warning results into the above-mentioned accident report template to generate an accident warning report.

[0071] The third step is to send the aforementioned accident warning report to the target terminal to perform the warning operation. This warning operation can involve sending an accident warning report to the vehicle driven by the target driver, or controlling the broadcasting equipment around the target driver's vehicle to broadcast the accident warning information.

[0072] The various embodiments of this disclosure have the following beneficial effects: the traffic accident early warning method under severe weather conditions according to some embodiments of this disclosure improves the driving safety of drivers. Specifically, the reason for the low driving safety of drivers is that when data analysis is performed using only single-dimensional data, the data content is not comprehensive enough, the impact of severe weather on driving is not considered, and a dynamic matching mechanism between driver profiles and regional risk characteristics cannot be established, resulting in biased early warning results and low driving safety of drivers. Based on this, the traffic accident early warning method under severe weather conditions according to some embodiments of this disclosure firstly, in response to receiving driving monitoring information for a target driver sent by a target terminal, obtains a multimodal dataset corresponding to the target driver. Thus, multi-dimensional data of the driver can be obtained. Secondly, the multimodal dataset is fused to generate a fused data matrix. Thus, data from different dimensions can be unified to the same dimension through data fusion. Then, driver feature extraction processing is performed on the fused data matrix to generate an extracted feature set. Thus, various data features of the multimodal data can be extracted. Afterwards, based on the extracted feature set, the behavioral ability of the target driver is evaluated to generate a driver behavior evaluation value. Thus, a driver ability evaluation can be generated to determine the driver's driving ability. Next, a traffic accident early warning model corresponding to the extracted feature set is selected from the traffic accident early warning model set as the target early warning model. The fused data matrix is ​​then input into the target early warning model to obtain the accident early warning result. This allows for the prediction of road traffic accidents. Finally, based on the driver behavior evaluation value and the accident early warning result, an accident early warning report is generated and sent to the target terminal for early warning operation. This completes the early warning of traffic accidents. Furthermore, by fusing and analyzing multimodal driver data, a dynamic matching mechanism between driver profiles and regional risk characteristics can be established, generating accurate early warning results and improving driver safety.

[0073] Further reference Figure 2 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a traffic accident warning device under severe weather conditions. These device embodiments are similar to... Figure 1 Corresponding to the method embodiments shown, this traffic accident warning device under severe weather conditions can be specifically applied to various electronic devices.

[0074] like Figure 2As shown, a traffic accident warning device 200 under severe weather conditions in some embodiments includes: an acquisition unit 201, a fusion unit 202, a feature extraction unit 203, an evaluation unit 204, a selection unit 205, and a generation unit 206. The acquisition unit 201 is configured to acquire a multimodal dataset corresponding to the target driver in response to receiving driving monitoring information for the target driver sent by the target terminal; the fusion unit 202 is configured to perform fusion processing on the multimodal dataset to generate a fused data matrix; the feature extraction unit 203 is configured to perform driver feature extraction processing on the fused data matrix to generate an extracted feature set; the evaluation unit 204 is configured to evaluate the behavior ability of the target driver based on the extracted feature set to generate a driver behavior evaluation value; the selection unit 205 is configured to select a traffic accident early warning model corresponding to the extracted feature set from the traffic accident early warning model set as the target early warning model, and input the fused data matrix into the target early warning model to obtain an accident early warning result; the generation unit 206 is configured to generate an accident early warning report based on the driver behavior evaluation value and the accident early warning result, and send the accident early warning report to the target terminal for early warning operation.

[0075] It is understandable that the various units and references recorded in the traffic accident warning device 200 under severe weather conditions are... Figure 1 The steps in the described method correspond to each other. Therefore, the operations, features, and beneficial effects described above for the method are also applicable to the traffic accident warning device 200 under severe weather conditions and the units contained therein, and will not be repeated here.

[0076] The following is for reference. Figure 3 This document illustrates a structural schematic of an electronic device 300 suitable for implementing some embodiments of the present disclosure. The electronic devices in some embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0077] like Figure 3As shown, the electronic device 300 may include a processing unit 301 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device 300. The processing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0078] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 3 Each box shown can represent a device or multiple devices as needed.

[0079] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from storage device 308, or installed from ROM 302. When the computer program is executed by processing device 301, it performs the functions defined in the methods of some embodiments of this disclosure.

[0080] It should be noted that, in some embodiments of this disclosure, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0081] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0082] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: in response to receiving driving monitoring information for a target driver sent by a target terminal, acquire a multimodal dataset corresponding to the target driver; perform fusion processing on the multimodal dataset to generate a fused data matrix; perform driver feature extraction processing on the fused data matrix to generate an extracted feature set; evaluate the target driver's behavioral ability based on the extracted feature set to generate a driver behavior evaluation value; select a traffic accident early warning model corresponding to the extracted feature set from a set of traffic accident early warning models as the target early warning model, and input the aforementioned fused data matrix into the target early warning model to obtain an accident early warning result; and generate an accident early warning report based on the driver behavior evaluation value and the accident early warning result, and send the accident early warning report to the target terminal for early warning operation.

[0083] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0084] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. 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 indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated 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 the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0085] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including an acquisition unit, a fusion unit, a feature extraction unit, an evaluation unit, a selection unit, and a generation unit. The names of these units do not necessarily limit the specific unit; for example, the acquisition unit may also be described as "a unit that acquires a multimodal dataset corresponding to the target driver in response to receiving driving monitoring information for the target driver sent by a target terminal."

[0086] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0087] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A method for early warning of traffic accidents under severe weather conditions, comprising: In response to receiving driving monitoring information for the target driver sent by the target terminal, a multimodal dataset corresponding to the target driver is obtained; The multimodal dataset is fused to generate a fused data matrix; The fused data matrix is ​​subjected to driver feature extraction processing to generate an extracted feature set; Based on the extracted feature set, the target driver's behavioral ability is evaluated to generate a driver behavior evaluation value; The step of evaluating the target driver's behavioral capabilities based on the extracted feature set to generate a driver behavior evaluation value includes: The extracted features in the extracted feature set are sorted by timeliness to generate an extracted feature sequence; Determine the weight feature value corresponding to each extracted feature in the extracted feature sequence to generate a weight feature sequence; Generate a driver behavior ability table; For each driver behavior ability in the driver behavior ability table, the following processing steps are performed: Determine at least one extractable feature corresponding to the driver's behavioral capabilities; Based on the at least one extracted feature, the driver's behavioral ability is predicted to generate a predicted ability value. Determine the scene type corresponding to the multimodal dataset; Based on the scenario type, the predicted capability value is corrected to generate a corrected capability value; Based on the generated correction capability values, an intervention strategy matrix corresponding to the target driver is generated, and the sum of the generated correction capability values ​​is determined as the driver behavior evaluation value. The traffic accident early warning model corresponding to the extracted feature set is selected from the traffic accident early warning model set as the target early warning model, and the fused data matrix is ​​input into the target early warning model to obtain the accident early warning result; The step of selecting a traffic accident early warning model corresponding to the extracted feature set from the traffic accident early warning model set as the target early warning model, and inputting the fused data matrix into the target early warning model to obtain the accident early warning result, includes: Based on the extracted feature set, the environmental feature information corresponding to the target driver is determined; Determine the similarity between the environmental feature information and each preset scene feature information in the preset scene feature information set to generate a scene similarity set; Based on the scene similarity set, target scene feature information is selected from the preset scene feature information set, and the traffic accident early warning model corresponding to the target scene feature information is used as the target early warning model. The fused data matrix is ​​input into the target early warning model to obtain the accident early warning result; Based on the driver behavior evaluation value and the accident warning result, an accident warning report is generated, and the accident warning report is sent to the target terminal for warning operation.

2. The method according to claim 1, wherein, The process of fusing the multimodal dataset to generate a fused data matrix includes: For each multimodal data point in the multimodal dataset, determine the confidence level corresponding to the multimodal data point to generate a data confidence level; Based on the generated confidence scores of each data point, the multimodal data in the multimodal dataset are fused to generate a fused data matrix.

3. The method according to claim 1, wherein, Before performing the fusion processing on the multimodal dataset to generate a fused data matrix, the method further includes: Each multimodal data in the multimodal dataset is cleaned to generate a cleaned dataset; The cleaned dataset is subjected to feature alignment processing to generate an aligned dataset, which serves as a multimodal dataset.

4. The method according to claim 1, wherein, The step of generating an accident warning report based on the driver behavior evaluation value and the accident warning result, and sending the accident warning report to the target terminal for warning operation, includes: Get the accident report template; The various correction capability values, the intervention strategy matrix, the driver behavior evaluation value, and the accident warning result are entered into the accident report template to generate an accident warning report; The accident warning report is sent to the target terminal for warning operation.

5. A traffic accident early warning device for severe weather, comprising: The acquisition unit is configured to acquire a multimodal dataset corresponding to the target driver in response to receiving driving monitoring information for the target driver sent by the target terminal; The fusion unit is configured to perform fusion processing on the multimodal dataset to generate a fused data matrix; The feature extraction unit is configured to perform driver feature extraction processing on the fused data matrix to generate an extracted feature set; An evaluation unit is configured to evaluate the behavioral capabilities of the target driver based on the extracted feature set, thereby generating a driver behavior evaluation value; the evaluation unit is further configured to: The extracted features in the extracted feature set are sorted by timeliness to generate an extracted feature sequence; Determine the weight feature value corresponding to each extracted feature in the extracted feature sequence to generate a weight feature sequence; Generate a driver behavior ability table; For each driver behavior ability in the driver behavior ability table, the following processing steps are performed: Determine at least one extractable feature corresponding to the driver's behavioral capabilities; Based on the at least one extracted feature, the driver's behavioral ability is predicted to generate a predicted ability value. Determine the scene type corresponding to the multimodal dataset; Based on the scenario type, the predicted capability value is corrected to generate a corrected capability value; Based on the generated correction capability values, an intervention strategy matrix corresponding to the target driver is generated, and the sum of the generated correction capability values ​​is determined as the driver behavior evaluation value. The selection unit is configured to select a traffic accident early warning model corresponding to the extracted feature set from the traffic accident early warning model set as the target early warning model, and to input the fused data matrix into the target early warning model to obtain the accident early warning result; the selection unit is further configured to: Based on the extracted feature set, the environmental feature information corresponding to the target driver is determined; Determine the similarity between the environmental feature information and each preset scene feature information in the preset scene feature information set to generate a scene similarity set; Based on the scene similarity set, target scene feature information is selected from the preset scene feature information set, and the traffic accident early warning model corresponding to the target scene feature information is used as the target early warning model. The fused data matrix is ​​input into the target early warning model to obtain the accident early warning result; The generation unit is configured to generate an accident warning report based on the driver behavior evaluation value and the accident warning result, and to send the accident warning report to the target terminal for warning operation.

6. An electronic device, comprising: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 4.

7. A computer-readable medium having a computer program stored thereon, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 4.

Images