Road risk behavior prediction method and device, computer equipment, readable storage medium and program product
Through intelligent monitoring equipment and target detection models, the trajectory of vulnerable road users is detected and tracked, and combined with the hazard trajectory prediction model, the problem of difficulty in detecting and blocking dangerous behaviors of vulnerable road users is solved, and accurate detection and risk prediction of these users are achieved, and traffic safety is improved.
Patent Information
- Application Number
- CN202510017120.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-06
- Publication Date
- 2025-05-02
AI Technical Summary
Existing autonomous driving technology is difficult to detect and predict dangerous behaviors that cannot be detected by vehicles that block vulnerable road users, resulting in possible traffic accidents.
The intelligent monitoring device obtains road video data in real time, and uses the target detection model to detect vulnerable road users. If they are not in the specified safe area, they will track trajectory tracking as the target of interest, and enter a hazard trajectory prediction model to obtain the trajectory risk level. If the risk level meets the preset conditions, the trajectory map, risk level and location information will be encapsulated and sent to the cloud, and finally the real-time trajectory map and risk level will be sent to the vehicle end within the predetermined distance of the position information.
It realizes accurate detection and risk prediction of vulnerable road users that cannot be detected by vehicle sensors, reduces the occurrence of traffic accidents and improves driving safety.
Smart Images
Figure CN119919897A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of autonomous driving technology, and in particular to a road risk behavior prediction method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Art
[0002] With the development of my country's transportation industry, the number of motor vehicles has increased dramatically, and the contradiction between vehicles and vulnerable road users such as pedestrians, bicycles, and electric vehicles has become increasingly prominent. The movement trajectory of vulnerable road users is highly uncertain. If their dangerous behaviors cannot be predicted in advance, such as "ghosting" caused by crossing the road, fighting on the street, and uncivilized behavior of ignoring traffic lights, it may cause obstacle avoidance failure and lead to huge traffic accidents. Therefore, predicting the dangerous behaviors of vulnerable road users can effectively reduce such traffic accidents and ensure driving safety.
[0003] Currently, the perception of intelligent driving systems is mostly through obstacle detection and tracking through data returned by sensors on the vehicle. However, due to the limitations of sensor perception range and performance, the sensors on the vehicle are unable to detect and predict trajectories of obscured vulnerable road users. Summary of the invention
[0004] Based on this, it is necessary to provide a road risk behavior prediction method, device, computer equipment, computer-readable storage medium and computer program product that can detect obstructed vulnerable road users that cannot be detected by vehicles in order to address the above technical problems.
[0005] In a first aspect, the present application provides a road risk behavior prediction method, which is applied to a road-side device, comprising:
[0006] Obtain road video data in real time through intelligent monitoring equipment;
[0007] Detecting vulnerable road users in the road video data by using a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane;
[0008] Tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, and inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level;
[0009] When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; to instruct the cloud device to receive the packaged data; to parse the packaged data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0010] In one embodiment, there are multiple targets of interest, each target of interest corresponds to a real-time trajectory map, and each real-time trajectory map corresponds to a trajectory risk level; when the trajectory risk level meets the preset reporting condition, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, including:
[0011] The real-time trajectory map of the target whose trajectory risk level meets the preset reporting conditions, the trajectory risk level corresponding to the real-time trajectory map of the target, and the position information of the target of interest corresponding to the real-time trajectory map of the target are encapsulated.
[0012] In one embodiment, the training of the dangerous trajectory prediction model includes:
[0013] Obtain motion videos of vulnerable road users, identify the motion videos through target tracking, and obtain motion trajectories of the vulnerable road users; label each motion trajectory, wherein the labels indicate the risk level of each motion trajectory, and obtain a risk data set; build a dangerous trajectory prediction model to be trained, and use the risk data set to train the dangerous trajectory prediction model to be trained, and obtain the dangerous trajectory prediction model.
[0014] In one embodiment, the process of determining whether the risk level meets the preset reporting conditions is as follows:
[0015] If the trajectory risk level is a normal trajectory, the preset reporting condition is not met; if the trajectory risk level has an abnormal trajectory, the preset reporting condition is met.
[0016] In a second aspect, the present application provides a road risk behavior prediction method, which is applied to a cloud device, including:
[0017] Receiving encapsulated data from a road-side device; the encapsulated data is obtained by the road-side device in real time through an intelligent monitoring device to obtain road video data; detecting vulnerable road users in the road video data through a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of a sidewalk or a non-motor vehicle lane; tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level; when the trajectory risk level meets a preset reporting condition, encapsulating the real-time trajectory map, the trajectory risk level and the location information of the target of interest;
[0018] Parsing the packaged data to obtain the real-time trajectory map, the risk level and the location information;
[0019] The real-time trajectory map and the risk level are sent to a vehicle terminal within a predetermined distance of the location information.
[0020] In one embodiment, the real-time trajectory map and the risk level are sent to the vehicle terminal within a predetermined distance of the location information.
[0021] Based on the location information, all vehicles within a predetermined distance of the location information are determined; and the real-time trajectory map and the risk level are distributed to the vehicles through the DDS protocol.
[0022] In a third aspect, the present application also provides a road risk behavior prediction device, which is applied to road-side equipment, including:
[0023] An acquisition module is used to acquire road video data in real time through intelligent monitoring equipment;
[0024] A detection module, configured to detect vulnerable road users in the road video data by using a target detection model, and to take vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; and the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane;
[0025] a risk prediction module, configured to track the trajectory of the target of interest in the road video data, obtain a real-time trajectory map of the target of interest, input the real-time trajectory map into a dangerous trajectory prediction model, and obtain a trajectory risk level;
[0026] The sending module is used to encapsulate the real-time trajectory map, the trajectory risk level and the location information of the target of interest when the trajectory risk level meets the preset reporting conditions, and send the encapsulated data to the cloud; to instruct the cloud to receive the encapsulated data; to parse the encapsulated data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0027] In a fourth aspect, the present application also provides a road risk behavior prediction device, which is applied to a cloud device, including:
[0028] A receiving module is used to receive encapsulated data from a road-side device; the encapsulated data is obtained by the road-side device in real time through an intelligent monitoring device to obtain road video data; vulnerable road users in the road video data are detected through a target detection model, and vulnerable road users who are not in a specified safety area are taken as targets of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of a sidewalk or a non-motor vehicle lane; the trajectory of the target of interest is tracked in the road video data to obtain a real-time trajectory map of the target of interest, and the real-time trajectory map is input into a dangerous trajectory prediction model to obtain a trajectory risk level; when the trajectory risk level meets a preset reporting condition, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are encapsulated to obtain;
[0029] A parsing module, used for parsing the encapsulated data to obtain the real-time trajectory map, the risk level and the location information;
[0030] A distribution module is used to send the real-time trajectory map and the risk level to the vehicle terminal within a predetermined distance of the location information.
[0031] In a fifth aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:
[0032] Obtain road video data in real time through intelligent monitoring equipment;
[0033] Detecting vulnerable road users in the road video data by using a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane;
[0034] Tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, and inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level;
[0035] When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; to instruct the cloud device to receive the packaged data; to parse the packaged data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0036] In a sixth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:
[0037] Obtain road video data in real time through intelligent monitoring equipment;
[0038] Detecting vulnerable road users in the road video data by using a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane;
[0039] Tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, and inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level;
[0040] When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; to instruct the cloud device to receive the packaged data; to parse the packaged data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0041] In a seventh aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:
[0042] Obtain road video data in real time through intelligent monitoring equipment;
[0043] Detecting vulnerable road users in the road video data by using a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane;
[0044] Tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, and inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level;
[0045] When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; to instruct the cloud device to receive the packaged data; to parse the packaged data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0046] The above-mentioned road risk behavior prediction method, device, computer equipment, computer-readable storage medium and computer program product obtain road video data in real time through intelligent monitoring equipment; detect vulnerable road users in the road video data through a target detection model, and take vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of a sidewalk or a non-motor vehicle lane; track the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, input the real-time trajectory map into a dangerous trajectory prediction model, and obtain a trajectory risk level; when the trajectory risk level meets a preset reporting condition, encapsulate the real-time trajectory map, the trajectory risk level and the location information of the target of interest, and send the encapsulated data to the cloud; to instruct the cloud device to receive the encapsulated data; parse the encapsulated data to obtain the real-time trajectory map, the risk level and the location information; send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0048] Figure 1A diagram showing an application environment of a road risk behavior prediction method in an embodiment;
[0049] Figure 2 A schematic diagram of a process of applying a road risk behavior prediction method to a roadside device in one embodiment;
[0050] Figure 3 A schematic diagram of a process of applying a road risk behavior prediction method to a cloud device in one embodiment;
[0051] Figure 4 A detailed flowchart of a road risk behavior prediction method in one embodiment;
[0052] Figure 5 It is a structural block diagram of a road risk behavior prediction device applied to a road-side device in one embodiment;
[0053] Figure 6 It is a structural block diagram of a road risk behavior prediction device applied to a cloud device in one embodiment;
[0054] Figure 7 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0055] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0056] The road risk behavior prediction method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown, the vehicle 102, the cloud device 104 and the roadside device 106 all communicate through the network. The data storage system can store the data that the cloud device 104 needs to process. The data storage system can be placed on the cloud or other network servers. The road-side device 106 obtains road video data in real time through intelligent monitoring equipment; detects vulnerable road users in the road video data through a target detection model, and takes vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized lane; tracks the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, inputs the real-time trajectory map into a dangerous trajectory prediction model, and obtains a trajectory risk level; when the trajectory risk level meets a preset reporting condition, encapsulates the real-time trajectory map, the trajectory risk level, and the location information of the target of interest, and sends the encapsulated data to the cloud device 104; to instruct the cloud device to receive the encapsulated data; parses the encapsulated data to obtain the real-time trajectory map, the risk level, and the location information; and sends the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information. The cloud device 104 is a cloud server that provides cloud computing services.
[0057] In an exemplary embodiment, Figure 2 As shown, a road risk behavior prediction method is provided, and the method is applied to Figure 1 The road end device 106 in FIG. 1 is taken as an example to illustrate, including the following steps 202 to 206. Among them:
[0058] Step 202: Acquire road video data in real time through intelligent monitoring equipment.
[0059] Among them, the intelligent monitoring equipment is arranged near the road-side equipment, such as next to the traffic lights. In addition to the intelligent monitoring equipment, the vehicle-side cameras near the road-side equipment also collect video data and send it to the road-side equipment through the cloud. The video data collected by the vehicle camera and the video data collected by the intelligent monitoring together constitute the road video data.
[0060] Exemplarily, the road video data is acquired in real time through the intelligent monitoring device, the road video data shot by the vehicle from the cloud device is received, and all the road video data is used as the road video data for identification.
[0061] Step 204, detecting vulnerable road users in the road video data by using a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane.
[0062] Among them, the designated safe area refers to the area where vulnerable road users should move. For example, pedestrians should cross the road at zebra crossings. Those beyond the zebra crossings are not within the designated safe area; bicycles and electric vehicles should travel on non-motorized vehicle lanes.
[0063] Exemplarily, vulnerable road users in road video data are detected through a target detection model, and pedestrians not on zebra crossings and bicycles and electric vehicles not on non-motorized vehicle lanes are taken as targets of interest.
[0064] Step 206 , tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level.
[0065] Optionally, risk levels include: safe trajectory, low risk trajectory, medium risk trajectory, and high risk trajectory.
[0066] Optionally, the dangerous trajectory prediction model is built based on 3D ResNet.
[0067] Exemplarily, the trajectory of the target of interest is tracked in the road video data to obtain real-time trajectory maps of all targets of interest, all real-time trajectory maps are input into the dangerous trajectory prediction model, and the trajectory risk level of each real-time trajectory map is output, including: safe trajectory, low-risk trajectory, medium-risk trajectory and high-risk trajectory.
[0068] Step 208, when the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; to instruct the cloud device to receive the packaged data; to parse the packaged data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0069] Among them, the preset reporting condition is that when the risk level of the trajectory is a low-risk trajectory, a medium-risk trajectory or a high-risk trajectory, it will be reported to the cloud device.
[0070] Optionally, the predetermined distance may be a distance of 100 meters near the location information.
[0071] Exemplarily, if the trajectory risk level is a low-risk trajectory, a medium-risk trajectory, or a high-risk trajectory, the real-time trajectory map, the trajectory risk level, and the location information of the target of interest are packaged, and the packaged data is sent to the cloud; to instruct the cloud to receive the packaged data; to parse the packaged data to obtain the real-time trajectory map, the risk level, and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0072] The above-mentioned road risk behavior prediction method, device, computer equipment, computer-readable storage medium and computer program product obtain road video data in real time through intelligent monitoring equipment; detect vulnerable road users in the road video data through the target detection model, and take vulnerable road users who are not in the specified safety area as the target of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of sidewalks or non-motorized vehicle lanes; track the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, input the real-time trajectory map into the dangerous trajectory prediction model, and obtain the trajectory risk level; when the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; to instruct the cloud device to receive the packaged data; parse the packaged data to obtain the real-time trajectory map, the risk level and the location information; send the real-time trajectory map and the risk level to the vehicle end within the predetermined distance of the location information. Vulnerable road users that cannot be detected by vehicle sensors can be accurately detected to prevent the occurrence of safety accidents.
[0073] In an exemplary embodiment, there are multiple targets of interest, each target of interest corresponds to a real-time trajectory map, and each real-time trajectory map corresponds to a trajectory risk level; when the trajectory risk level meets the preset reporting condition, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, including:
[0074] The real-time trajectory map of the target whose trajectory risk level meets the preset reporting conditions, the trajectory risk level corresponding to the real-time trajectory map of the target, and the position information of the target of interest corresponding to the real-time trajectory map of the target are encapsulated.
[0075] When there are multiple targets of interest in a video, a real-time trajectory map is generated based on each target of interest, and a trajectory risk level is generated based on each real-time trajectory map. When there are multiple intelligent monitoring devices and multiple vehicle sensors, multiple videos will be generated, and the above operation will be performed on each video.
[0076] Exemplarily, a real-time trajectory map of a trajectory risk level of a low-risk trajectory, a medium-risk trajectory, or a high-risk trajectory is used as a target real-time trajectory map, and the target real-time trajectory map, the trajectory risk level corresponding to the target real-time trajectory map, and the location information of the target of interest corresponding to the target real-time trajectory map are encapsulated.
[0077] In this embodiment, by encapsulating the real-time trajectory graphs and related information corresponding to the low-risk trajectory, the medium-risk trajectory, and the high-risk trajectory, data can be conveniently transmitted.
[0078] In an exemplary embodiment, the training of the dangerous trajectory prediction model includes:
[0079] Obtain motion videos of vulnerable road users, identify the motion videos through target tracking, and obtain motion trajectories of the vulnerable road users; label each motion trajectory, wherein the labels indicate the risk level of each motion trajectory, and obtain a risk data set; build a dangerous trajectory prediction model to be trained, and use the risk data set to train the dangerous trajectory prediction model to be trained, and obtain the dangerous trajectory prediction model.
[0080] Exemplarily, a large number of motion videos of vulnerable road users are obtained, the motion videos are identified through a target tracking model, and motion trajectory maps of vulnerable road users are obtained. Each motion trajectory map is labeled with the following labels: safe trajectory, low-risk trajectory, medium-risk trajectory and high-risk trajectory; the motion trajectory maps and corresponding labels are divided into a training set and a test set, thereby obtaining a risk data set; a dangerous trajectory prediction model to be trained is built, the dangerous trajectory prediction model to be trained is trained using the training set of the risk data set, and the progress of the training is verified using the test set, and finally a trained dangerous trajectory prediction model is obtained.
[0081] In this embodiment, by establishing a data set and building a model for training, a dangerous trajectory prediction model that accurately identifies the trajectory risk level can be obtained.
[0082] In an exemplary embodiment, the process of determining whether the risk level meets the preset reporting conditions is as follows:
[0083] If the trajectory risk level is a normal trajectory, the preset reporting condition is not met; if the trajectory risk level has an abnormal trajectory, the preset reporting condition is met.
[0084] Exemplarily, if all trajectory risk levels are normal trajectories, the preset reporting condition is not met; if there are abnormal trajectories in the trajectory risk level, that is, there are low-risk trajectories, medium-risk trajectories or high-risk trajectories, the reporting condition is met.
[0085] In this embodiment, when the risk levels of the trajectories of vulnerable road users are all safe trajectories, no reporting is performed, thereby saving the amount of transmitted data and alleviating the pressure on network transmission.
[0086] In an exemplary embodiment, Figure 3 As shown, a road risk behavior prediction method is provided, and the method is applied to Figure 1 Taking the cloud device 104 in the example as an example, the following steps 302 to 306 are included. Among them:
[0087] Step 302, receiving encapsulated data from a road-side device; the encapsulated data is obtained by the road-side device in real time through an intelligent monitoring device to obtain road video data; vulnerable road users in the road video data are detected through a target detection model, and vulnerable road users who are not in a specified safety area are taken as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized lane; the trajectory of the target of interest is tracked in the road video data to obtain a real-time trajectory map of the target of interest, and the real-time trajectory map is input into a dangerous trajectory prediction model to obtain a trajectory risk level; when the trajectory risk level meets a preset reporting condition, the real-time trajectory map, the trajectory risk level and the position information of the target of interest are encapsulated to obtain.
[0088] Among them, the encapsulated data includes the real-time trajectory map, corresponding trajectory risk level and location information of all vulnerable road users whose trajectory risk level is above low-risk trajectory.
[0089] Step 304, parsing the packaged data to obtain the real-time trajectory map, the risk level and the location information.
[0090] Among them, the real-time trajectory map is the real-time trajectory map of all vulnerable road users with trajectory risk levels above low-risk trajectories, the risk level is the risk level of the real-time trajectory map of all vulnerable road users with trajectory risk levels above low-risk trajectories, and the location information is the location information of all vulnerable road users with trajectory risk levels above low-risk trajectories.
[0091] Step 306: Send the real-time trajectory map and the risk level to the vehicle terminal within a predetermined distance of the location information.
[0092] Optionally, the vehicle terminals within a predetermined distance of the location information are all vehicle terminals within 100 meters of the location information.
[0093] Exemplarily, the real-time trajectory map and risk level are sent to all vehicles within 100 meters of the location of the vulnerable road user.
[0094] The above-mentioned road risk behavior prediction method, device, computer equipment, computer-readable storage medium and computer program product receive encapsulated data from a road-side device; the encapsulated data is obtained by the road-side device in real time through intelligent monitoring equipment to obtain road video data; the vulnerable road users in the road video data are detected by the target detection model, and the vulnerable road users who are not in the specified safety area are taken as the target of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of a sidewalk or a non-motor vehicle lane; the trajectory of the target of interest is tracked in the road video data to obtain a real-time trajectory map of the target of interest, and the real-time trajectory map is input into the dangerous trajectory prediction model to obtain the trajectory risk level; when the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are encapsulated; the encapsulated data is parsed to obtain the real-time trajectory map, the risk level and the location information; the real-time trajectory map and the risk level are sent to the vehicle end within the predetermined distance of the location information. Vulnerable road users that cannot be detected by vehicle sensors can be accurately detected to prevent the occurrence of safety accidents.
[0095] In an exemplary embodiment, the real-time trajectory map and the risk level are sent to a vehicle terminal within a predetermined distance of the location information.
[0096] Based on the location information, all vehicles within a predetermined distance of the location information are determined; and the real-time trajectory map and the risk level are distributed to the vehicles through the DDS protocol.
[0097] Among them, the DDS protocol (Data Distribution Service) is designed to provide reliable, scalable and flexible data communication, and is widely used in systems that require real-time data exchange.
[0098] Exemplarily, based on the location information, all vehicles within 100 meters of the location information are determined, and the real-time trajectory map and risk level are sent to all vehicles through the DDS protocol.
[0099] In this embodiment, by sending the vulnerable road user trajectory with an abnormal danger trajectory level to the vehicle 100 meters near the corresponding vulnerable road user, the vehicle can be warned in advance during automatic driving and take corresponding measures.
[0100] In an exemplary embodiment, Figure 4As shown, the road-side device acquires in real time the road video data from the vehicle sensor sent by the intelligent monitoring device and the cloud; detects the vulnerable road users in the road video data through the target detection model, and takes the vulnerable road users who are not in the specified safety area as the target of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of a sidewalk or a non-motor vehicle lane; tracks the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, inputs the real-time trajectory map into the dangerous trajectory prediction model, and obtains the trajectory risk level; when the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; the cloud device receives the packaged data from the road-side device; parses the packaged data to obtain the real-time trajectory map, the risk level and the location information; and sends the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0101] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0102] In an exemplary embodiment, Figure 5 As shown, a road risk behavior prediction device is provided, which is applied to road-side equipment, including: an acquisition module 501, a detection module 502, a risk prediction module 503 and a sending module 504, wherein:
[0103] An acquisition module is used to acquire road video data in real time through intelligent monitoring equipment;
[0104] A detection module, configured to detect vulnerable road users in the road video data by using a target detection model, and to take vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; and the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane;
[0105] a risk prediction module, configured to track the trajectory of the target of interest in the road video data, obtain a real-time trajectory map of the target of interest, input the real-time trajectory map into a dangerous trajectory prediction model, and obtain a trajectory risk level;
[0106] The sending module is used to encapsulate the real-time trajectory map, the trajectory risk level and the location information of the target of interest when the trajectory risk level meets the preset reporting conditions, and send the encapsulated data to the cloud; to instruct the cloud to receive the encapsulated data; to parse the encapsulated data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0107] In one embodiment, the sending module is further used to:
[0108] There are multiple targets of interest, each target of interest corresponds to a real-time trajectory map, and each real-time trajectory map corresponds to a trajectory risk level; when the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are encapsulated, including: encapsulating the target real-time trajectory map whose trajectory risk level meets the preset reporting conditions, the trajectory risk level corresponding to the target real-time trajectory map, and the location information of the target of interest corresponding to the target real-time trajectory map.
[0109] In one embodiment, the risk prediction module is further used to:
[0110] Obtain motion videos of vulnerable road users, identify the motion videos through target tracking, and obtain motion trajectories of the vulnerable road users; label each motion trajectory, wherein the labels indicate the risk level of each motion trajectory, and obtain a risk data set; build a dangerous trajectory prediction model to be trained, and use the risk data set to train the dangerous trajectory prediction model to be trained, and obtain the dangerous trajectory prediction model.
[0111] In one embodiment, the sending module is further used to:
[0112] If the trajectory risk level is a normal trajectory, the preset reporting condition is not met; if the trajectory risk level has an abnormal trajectory, the preset reporting condition is met.
[0113] In an exemplary embodiment, Figure 6 As shown, a road risk behavior prediction device is provided, which is applied to a cloud device, including: a receiving module 601, a parsing module 602 and a distribution module 603, wherein:
[0114] A receiving module is used to receive encapsulated data from a road-side device; the encapsulated data is obtained by the road-side device in real time through an intelligent monitoring device to obtain road video data; vulnerable road users in the road video data are detected through a target detection model, and vulnerable road users who are not in a specified safety area are taken as targets of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of a sidewalk or a non-motor vehicle lane; the trajectory of the target of interest is tracked in the road video data to obtain a real-time trajectory map of the target of interest, and the real-time trajectory map is input into a dangerous trajectory prediction model to obtain a trajectory risk level; when the trajectory risk level meets a preset reporting condition, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are encapsulated to obtain;
[0115] A parsing module, used for parsing the encapsulated data to obtain the real-time trajectory map, the risk level and the location information;
[0116] A distribution module is used to send the real-time trajectory map and the risk level to the vehicle terminal within a predetermined distance of the location information.
[0117] In one embodiment, the distribution module is further configured to:
[0118] Based on the location information, all vehicles within a predetermined distance of the location information are determined; and the real-time trajectory map and the risk level are distributed to the vehicles through the DDS protocol.
[0119] Each module in the above-mentioned road risk behavior prediction device can be implemented in whole or in part by software, hardware or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.
[0120] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 7As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store road video data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a road risk behavior prediction method is implemented.
[0121] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0122] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:
[0123] Obtain road video data in real time through intelligent monitoring equipment;
[0124] Detecting vulnerable road users in the road video data by using a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane;
[0125] Tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, and inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level;
[0126] When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; to instruct the cloud device to receive the packaged data; to parse the packaged data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0127] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0128] When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are encapsulated, including: encapsulating the real-time trajectory map of the target whose trajectory risk level meets the preset reporting conditions, the trajectory risk level corresponding to the target real-time trajectory map, and the location information of the target of interest corresponding to the target real-time trajectory map.
[0129] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0130] Obtain motion videos of vulnerable road users, identify the motion videos through target tracking, and obtain motion trajectories of the vulnerable road users; label each motion trajectory, wherein the labels indicate the risk level of each motion trajectory, and obtain a risk data set; build a dangerous trajectory prediction model to be trained, and use the risk data set to train the dangerous trajectory prediction model to be trained, and obtain the dangerous trajectory prediction model.
[0131] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0132] If the trajectory risk level is a normal trajectory, the preset reporting condition is not met; if the trajectory risk level has an abnormal trajectory, the preset reporting condition is met.
[0133] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:
[0134] Receiving encapsulated data from a road-side device; the encapsulated data is obtained by the road-side device in real time through an intelligent monitoring device to obtain road video data; detecting vulnerable road users in the road video data through a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of a sidewalk or a non-motor vehicle lane; tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level; when the trajectory risk level meets a preset reporting condition, encapsulating the real-time trajectory map, the trajectory risk level and the location information of the target of interest;
[0135] Parsing the packaged data to obtain the real-time trajectory map, the risk level and the location information;
[0136] The real-time trajectory map and the risk level are sent to a vehicle terminal within a predetermined distance of the location information.
[0137] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0138] Based on the location information, all vehicles within a predetermined distance of the location information are determined; and the real-time trajectory map and the risk level are distributed to the vehicles through the DDS protocol.
[0139] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:
[0140] Obtain road video data in real time through intelligent monitoring equipment;
[0141] Detecting vulnerable road users in the road video data by using a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane;
[0142] Tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, and inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level;
[0143] When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; to instruct the cloud device to receive the packaged data; to parse the packaged data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0144] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0145] When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are encapsulated, including: encapsulating the real-time trajectory map of the target whose trajectory risk level meets the preset reporting conditions, the trajectory risk level corresponding to the target real-time trajectory map, and the location information of the target of interest corresponding to the target real-time trajectory map.
[0146] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0147] Obtain motion videos of vulnerable road users, identify the motion videos through target tracking, and obtain motion trajectories of the vulnerable road users; label each motion trajectory, wherein the labels indicate the risk level of each motion trajectory, and obtain a risk data set; build a dangerous trajectory prediction model to be trained, and use the risk data set to train the dangerous trajectory prediction model to be trained, and obtain the dangerous trajectory prediction model.
[0148] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0149] If the trajectory risk level is a normal trajectory, the preset reporting condition is not met; if the trajectory risk level has an abnormal trajectory, the preset reporting condition is met.
[0150] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:
[0151] Receiving encapsulated data from a road-side device; the encapsulated data is obtained by the road-side device in real time through an intelligent monitoring device to obtain road video data; detecting vulnerable road users in the road video data through a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of a sidewalk or a non-motor vehicle lane; tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level; when the trajectory risk level meets a preset reporting condition, encapsulating the real-time trajectory map, the trajectory risk level and the location information of the target of interest;
[0152] Parsing the packaged data to obtain the real-time trajectory map, the risk level and the location information;
[0153] The real-time trajectory map and the risk level are sent to a vehicle terminal within a predetermined distance of the location information.
[0154] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0155] Based on the location information, all vehicles within a predetermined distance of the location information are determined; and the real-time trajectory map and the risk level are distributed to the vehicles through the DDS protocol.
[0156] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:
[0157] Obtain road video data in real time through intelligent monitoring equipment;
[0158] Detecting vulnerable road users in the road video data by using a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane;
[0159] Tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, and inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level;
[0160] When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; to instruct the cloud device to receive the packaged data; to parse the packaged data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
[0161] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0162] When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are encapsulated, including: encapsulating the real-time trajectory map of the target whose trajectory risk level meets the preset reporting conditions, the trajectory risk level corresponding to the target real-time trajectory map, and the location information of the target of interest corresponding to the target real-time trajectory map.
[0163] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0164] Obtain motion videos of vulnerable road users, identify the motion videos through target tracking, and obtain motion trajectories of the vulnerable road users; label each motion trajectory, wherein the labels indicate the risk level of each motion trajectory, and obtain a risk data set; build a dangerous trajectory prediction model to be trained, and use the risk data set to train the dangerous trajectory prediction model to be trained, and obtain the dangerous trajectory prediction model.
[0165] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0166] If the trajectory risk level is a normal trajectory, the preset reporting condition is not met; if the trajectory risk level has an abnormal trajectory, the preset reporting condition is met.
[0167] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:
[0168] Receiving encapsulated data from a road-side device; the encapsulated data is obtained by the road-side device in real time through an intelligent monitoring device to obtain road video data; detecting vulnerable road users in the road video data through a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of a sidewalk or a non-motor vehicle lane; tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level; when the trajectory risk level meets a preset reporting condition, encapsulating the real-time trajectory map, the trajectory risk level and the location information of the target of interest;
[0169] Parsing the packaged data to obtain the real-time trajectory map, the risk level and the location information;
[0170] The real-time trajectory map and the risk level are sent to a vehicle terminal within a predetermined distance of the location information.
[0171] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0172] Based on the location information, all vehicles within a predetermined distance of the location information are determined; and the real-time trajectory map and the risk level are distributed to the vehicles through the DDS protocol.
[0173] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.
[0174] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0175] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A road risk behavior prediction method, characterized in that: Applied to roadside equipment, the method comprises: Obtain road video data in real time through intelligent monitoring equipment; Detecting vulnerable road users in the road video data by using a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane; Tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, and inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level; When the trajectory risk level meets the preset reporting conditions, the real-time trajectory map, the trajectory risk level and the location information of the target of interest are packaged, and the packaged data is sent to the cloud device; to instruct the cloud device to receive the packaged data; to parse the packaged data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
2. The method according to claim 1, characterized in that There are multiple targets of interest, each target of interest corresponds to a real-time trajectory map, and each real-time trajectory map corresponds to a trajectory risk level; When the trajectory risk level meets a preset reporting condition, encapsulating the real-time trajectory map, the trajectory risk level and the location information of the target of interest includes: The real-time trajectory map of the target whose trajectory risk level meets the preset reporting conditions, the trajectory risk level corresponding to the real-time trajectory map of the target, and the position information of the target of interest corresponding to the real-time trajectory map of the target are encapsulated.
3. The method according to claim 1, characterized in that The training of the dangerous trajectory prediction model includes: Acquire a motion video of a vulnerable road user, identify the motion video through target tracking, and obtain a motion trajectory of the vulnerable road user; Label each motion trajectory, wherein the label indicates the risk level of each motion trajectory, to obtain a risk data set; A dangerous trajectory prediction model to be trained is constructed, and the dangerous trajectory prediction model to be trained is trained using the risk data set to obtain the dangerous trajectory prediction model.
4. The method according to claim 1, characterized in that: The process of determining whether the risk level meets the preset reporting conditions is as follows: If the trajectory risk level is a normal trajectory, the preset reporting condition is not met; If there is an abnormal trajectory within the trajectory risk level, the preset reporting conditions are met.
5. A road risk behavior prediction method, characterized in that: Applied to a cloud device, the method includes: Receiving encapsulated data from a road-side device; the encapsulated data is obtained by the road-side device in real time through an intelligent monitoring device to obtain road video data; detecting vulnerable road users in the road video data through a target detection model, and taking vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motor vehicles; the specified safety area includes at least one of a sidewalk or a non-motor vehicle lane; tracking the trajectory of the target of interest in the road video data to obtain a real-time trajectory map of the target of interest, inputting the real-time trajectory map into a dangerous trajectory prediction model to obtain a trajectory risk level; when the trajectory risk level meets a preset reporting condition, encapsulating the real-time trajectory map, the trajectory risk level and the location information of the target of interest; Parsing the packaged data to obtain the real-time trajectory map, the risk level and the location information; The real-time trajectory map and the risk level are sent to a vehicle terminal within a predetermined distance of the location information.
6. The method according to claim 5, characterized in that The real-time trajectory map and the risk level are sent to a vehicle terminal within a predetermined distance of the location information, comprising: Based on the location information, determining all vehicles within a predetermined distance of the location information; The real-time trajectory map and the risk level are distributed to the vehicle through the DDS protocol.
7. A road risk behavior prediction device, characterized in that: Applied to roadside equipment, the device comprises: An acquisition module is used to acquire road video data in real time through intelligent monitoring equipment; A detection module, configured to detect vulnerable road users in the road video data by using a target detection model, and to take vulnerable road users who are not in a specified safety area as targets of interest; the vulnerable road users include at least one of pedestrians or non-motorized vehicles; and the specified safety area includes at least one of a sidewalk or a non-motorized vehicle lane; a risk prediction module, configured to track the trajectory of the target of interest in the road video data, obtain a real-time trajectory map of the target of interest, input the real-time trajectory map into a dangerous trajectory prediction model, and obtain a trajectory risk level; The sending module is used to encapsulate the real-time trajectory map, the trajectory risk level and the location information of the target of interest when the trajectory risk level meets the preset reporting conditions, and send the encapsulated data to the cloud; to instruct the cloud to receive the encapsulated data; to parse the encapsulated data to obtain the real-time trajectory map, the risk level and the location information; and to send the real-time trajectory map and the risk level to the vehicle end within a predetermined distance of the location information.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.