Vehicle behavior identification method and system

Abstract

The invention provides a vehicle behavior recognition method and system, and belongs to the technical field of vehicle behavior recognition, and the method comprises the steps: constructing a target vehicle behavior portrait based on the multi-source monitoring data of a target vehicle; and generating information board information corresponding to the target vehicle based on the target vehicle behavior portrait and the road environment monitoring data. And in response to the fact that the target vehicle arrives at the information board matching position, displaying the information board information. According to the vehicle behavior identification method and system provided by the invention, the identification accuracy of vehicle behaviors can be improved, and the road service level is further improved.

Description

Technical Field

[0001] The present disclosure belongs to the technical field of vehicle behavior recognition, and more specifically, to a vehicle behavior recognition method and system. Background Art

[0002] In recent years, my country's highway informatization has achieved leapfrog development, and a variety of vehicle data monitoring systems have been established through various technical means. These systems can provide vehicle monitoring data. However, the existing technology for identifying or detecting vehicle behavior lacks efficient use of these monitoring data, and there is an inaccurate problem in identifying vehicle behavior.

[0003] At the same time, the public travel information service system on highways has been continuously improved, with an increase in service channels such as APPs and WeChat public accounts. However, there are prominent problems with the variable information boards on road sections. The content they publish is of a broadcast nature, and the content is mostly triggered manually. The level of intelligence is low, making it difficult to meet the society's demand for high-quality service levels. Summary of the invention

[0004] The purpose of the present disclosure is to provide a vehicle behavior recognition method and system to improve the recognition accuracy of vehicle behavior and thus improve the road service level.

[0005] A first aspect of an embodiment of the present disclosure provides a vehicle behavior recognition method, comprising: Build a target vehicle behavior profile based on multi-source monitoring data of the target vehicle; Generate information board information corresponding to the target vehicle based on the target vehicle behavior portrait and road environment monitoring data; In response to recognizing that the target vehicle has arrived at the information board matching position, the information board information is displayed.

[0006] A second aspect of the embodiments of the present disclosure provides a vehicle behavior recognition system, including: The vehicle behavior profiling module is used to construct a target vehicle behavior profile based on the multi-source monitoring data of the target vehicle; An information board module, used to generate information board information corresponding to the target vehicle based on the target vehicle behavior portrait and road environment monitoring data; The display module is used to display the information on the information board in response to identifying that the target vehicle has arrived at the matching position of the information board.

[0007] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the above-mentioned vehicle behavior recognition method when executing the computer program.

[0008] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned vehicle behavior recognition method are implemented.

[0009] The vehicle behavior recognition method and system provided by the embodiments of the present disclosure have the following beneficial effects: The disclosed embodiment integrates multi-source monitoring data to construct a target vehicle behavior profile, which can comprehensively and continuously monitor key vehicle information, accurately analyze vehicle driving behavior, provide a basis for subsequent operations, and improve the accuracy of vehicle behavior recognition. The disclosed embodiment combines vehicle behavior profiles with road environment monitoring data to generate information board information, and through rule engines and natural language generation technology, provides vehicles with personalized, accompanying information push and safe driving reminders, which improves the intelligence and humanity of highway management and effectively promotes the overall improvement of road service levels. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 A flow chart of a vehicle behavior recognition method provided by an embodiment of the present disclosure; Figure 2 A structural block diagram of a vehicle behavior recognition system provided by an embodiment of the present disclosure; Figure 3 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0012] In the following description, specific details such as specific system structures and technologies are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present disclosure. However, it should be clear to those skilled in the art that the present disclosure may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obstructing the description of the present disclosure with unnecessary details.

[0013] In order to make the purpose, technical solutions and advantages of the present disclosure more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0014] Please refer to Figure 1 , Figure 1A flowchart of a vehicle behavior recognition method provided in one embodiment of the present disclosure, the method may include S101 to S103.

[0015] S101: Construct a target vehicle behavior profile based on multi-source monitoring data of the target vehicle.

[0016] In this embodiment, the target vehicle refers to a vehicle on the road to be monitored and analyzed. Multi-source monitoring data refers to data from different monitoring devices or systems. Multi-source monitoring data may include road checkpoint video data, ETC gantry video data, radar detection data, and road environment monitoring data.

[0017] For example, data fusion analysis is performed on the perception detection data such as radar and video installed on the highway and the vehicle information collected by the ETC toll gantries to form a global continuous monitoring of key information such as license plate, model, location, speed, etc. of vehicles traveling on the highway (i.e., target vehicles). Based on the global dynamic information of the target vehicle on the highway section, real-time monitoring and analysis of vehicle driving behavior, such as normal, speeding, slow speed, and parking, is performed to form a behavior profile of the target vehicle.

[0018] For example, by making full use of and integrating the vehicle perception means and data of multiple data monitoring devices or systems on highways, a dynamic driving behavior portrait of vehicles is constructed. Based on edge computing, AI video detection, multi-source fusion, big data analysis, dynamic triggering of information boards and other technologies, full-domain dynamic monitoring of target vehicles on the road section is achieved, and the personalized, accompanying information push and service capabilities of variable information boards for passing vehicles are improved, and the holographic and intelligent level of highway vehicle traffic monitoring and information services is improved.

[0019] S102: Generate information board information corresponding to the target vehicle based on the target vehicle behavior portrait and road environment monitoring data.

[0020] In this embodiment, the target vehicle behavior profile refers to a comprehensive description of the driving characteristics of the target vehicle. The target vehicle behavior profile may include a record of the vehicle behavior characteristics and annotated behavior tags.

[0021] For example, a convolutional neural network (CNN) can be used to detect and track target vehicles, extract the location and trajectory information of the vehicle from the collected video data, and calculate the location and time difference between consecutive frames to obtain the speed and acceleration of the vehicle. For radar data, a signal processing algorithm is used to parse out the precise speed, distance and relative position information.

[0022] Rules are set based on traffic regulations and experience, and rules engines (such as Drools) are used to convert extracted behavior features into behavior labels. For example, if the speed of a vehicle on a certain road section exceeds the speed limit of that road section through calculation, the rule engine will mark it as "speeding"; based on the vehicle trajectory and lane information, if the vehicle frequently crosses the lane line and does not conform to the normal lane change logic, it can be marked as "illegal lane change".

[0023] In this embodiment, the road environment monitoring data may include parameters such as weather conditions, road conditions, and traffic flow, and these road environment monitoring data reflect the external environment state of the vehicle. The information board information is the content displayed on the variable information board on the highway. For example, the information board information may include safe driving tips, service facility guidance, and traffic condition notifications.

[0024] Exemplarily, a rule engine can be established, which stores information templates of information boards corresponding to different combinations of vehicle behaviors and road environments. The target vehicle behavior portrait and road environment monitoring data are obtained through the data interface, and input into the rule engine to match the appropriate information board information template. Then, the natural language generation technology is used to fill the content of the information template, generate the final information board information, and transmit it to the corresponding variable information board device through the communication network for display. When the target vehicle passes through some variable information board devices that match the position, the variable information board device can dynamically display personalized accompanying information services and safe driving reminders for the target vehicle.

[0025] S103: In response to identifying that the target vehicle has arrived at the information board matching position, displaying the information board information.

[0026] In this embodiment, the information board matching position refers to the position of the variable information board device pre-set on the highway. There can be multiple information board matching positions.

[0027] For example, the location of the target vehicle is continuously tracked through positioning technology and compared with the matching location of the information board. When the vehicle reaches the matching location, the corresponding information board information is extracted from the database according to the pre-set rules and logic, and sent to the variable information board device. After receiving the information on the information board, the variable information board device displays the information on the information board to the driver according to the set display format and duration, thereby realizing accurate information service for the target vehicle.

[0028] From the above, it can be concluded that the present disclosure can obtain multi-source monitoring data for a specific road section, improve the accuracy of vehicle behavior recognition, and help to deeply analyze the behavior patterns of vehicles in different road sections. The target road is divided into multiple second road sections according to the matching position of the information board to avoid repeated calculations and realize refined management of the target road.

[0029] Based on vehicle behavior portraits and road environment monitoring data, the disclosed embodiment uses rule engines and natural language generation technology to generate personalized information board information for target vehicles, provide accompanying information services and safe driving reminders, and meet the needs of different vehicles in different environments. It realizes accurate information services for target vehicles and improves the holographic and intelligent level of highway information services.

[0030] In one embodiment of the present disclosure, the multi-source monitoring data includes road checkpoint video data, ETC gantry video data and radar detection data.

[0031] Build a target vehicle behavior profile based on the multi-source monitoring data of the target vehicle, including: The road checkpoint video data and the ETC gantry video data are sequentially spliced ​​based on the driving time sequence of the target vehicle to obtain the spliced ​​target video data.

[0032] A first vehicle behavior profile of the target vehicle is determined based on the target video data.

[0033] A second vehicle behavior profile of the target vehicle is determined based on the radar detection data.

[0034] The second vehicle behavior profile is embedded into the first vehicle behavior profile based on the driving time sequence and driving position information of the target vehicle to obtain the target vehicle behavior profile.

[0035] In this embodiment, determining a first vehicle behavior portrait of a target vehicle based on target video data includes: Extract the timestamp information and vehicle information of each frame image in the target video data.

[0036] Based on the vehicle information of each frame image, the vehicle position information, vehicle appearance features and vehicle behavior information of the target vehicle in the frame image are determined.

[0037] A first vehicle behavior portrait of the target vehicle is constructed based on the timestamp information of each frame image in the target video data, the vehicle position information, the vehicle appearance features and the vehicle behavior information of the target vehicle.

[0038] In this embodiment, the road checkpoint video data includes video data captured by checkpoint cameras set at key locations such as road entrances, exits, and important road intersections, and the road checkpoint video data includes images of vehicles passing through the checkpoints. ETC gantry video data includes video data collected at the ETC toll gantry on the highway. Radar detection data includes data obtained by detecting vehicles using radar equipment.

[0039] The first vehicle behavior profile is constructed based on video data, and may include the target vehicle's driving trajectory, dwell time at different checkpoints and gantries, appearance features, behavior features, etc. The second vehicle behavior profile is constructed based on radar data, and may include the vehicle's precise speed, acceleration change curve, distance maintenance, and other dynamic driving features.

[0040] The driving time sequence refers to the time sequence of each monitoring point during the target vehicle's driving on the road. It is sorted based on the timestamp and is used to associate the data of different monitoring points. The driving position information refers to the geographical location of the target vehicle on the road, which is determined by the known position of the bayonet, gantry and video image analysis or radar positioning, such as how many meters away from a service area, the specific location on a certain road section, etc.

[0041] The vehicle information of each frame includes the information of all vehicles in the frame, such as the license plate information and appearance information of each vehicle. The target vehicle can be identified based on the license plate information or appearance information. The vehicle position information refers to the relative position in the target vehicle video image and the actual geographical location relative to the monitoring device. The vehicle appearance features can include the license plate, model, color, etc.

[0042] Vehicle behavior information may include whether the vehicle is stationary (parked), driving, and whether the vehicle has any illegal behavior (such as driving in the wrong direction, illegal lane change). This can be determined by comparing and analyzing the vehicle's moving direction and speed changes through continuous frame analysis.

[0043] In this embodiment, the road checkpoint video data and the ETC gantry video data are sequentially spliced ​​based on the driving time sequence of the target vehicle to obtain the spliced ​​target video data, including: Use open source video processing libraries such as FFmpeg to extract video metadata, which includes timestamp information. Time-sort the road checkpoint video data and ETC gantry video data based on the timestamp information. Use video editing tools or custom video stitching programs to stitch the sorted video data into a continuous video file in chronological order to form the stitched target video data.

[0044] In this embodiment, the target detection algorithm in deep learning (such as YOLOv5) can be used to process each frame of the image to detect the position and bounding box of the vehicle, and the position information of the vehicle can be calculated according to the frame resolution and the detection result. The image classification technology is used to identify the appearance features of the vehicle, and the position changes of the previous and next frames are analyzed by algorithms such as the optical flow method to determine whether the vehicle is accelerating, decelerating, turning or stopping, etc., and finally the first vehicle behavior portrait is constructed.

[0045] Parse the radar data to obtain data such as the speed and distance of the vehicle. According to the installation position of the radar and the known road parameters, convert the distance and speed information into the position information of the vehicle. According to the time series of the speed data, calculate the acceleration of the vehicle using the difference method, thereby constructing the second vehicle behavior portrait.

[0046] Align the data of the first vehicle behavior portrait and the second vehicle behavior portrait according to the timestamp and position information. For the information with the same timestamp, add the accurate speed, distance, etc. information in the second vehicle behavior portrait to the first vehicle behavior portrait; for the data with similar but not exactly the same timestamps, linear interpolation or other smoothing algorithms can be used to supplement the radar data information into the first vehicle behavior portrait. Use the Pandas library in Python to store the two portraits as data frames, merge and fill the data according to the timestamp and position, and finally obtain the complete target vehicle behavior portrait.

[0047] Exemplarily, on a certain section of the highway, a truck is driving from the entrance to the exit. The bayonet camera at the road entrance captured the video footage of the truck entering the highway, recorded the timestamp as 9:00:00, extracted the license plate number of the truck as "Ji A12345", the vehicle type as a van, the body color as blue, and it was driving normally at the bayonet without stopping.

[0048] As the truck travels, the ETC gantry video data shows that the truck passes a certain ETC gantry at 9:15:00. At the same time, the radar detection data indicates that the speed of the truck is 80 km / h at this time, and the vehicle distance remains within the safe range, and the acceleration is stable.

[0049] By splicing the bayonet video and the ETC gantry video in chronological order, analyze and construct the first vehicle behavior portrait, including information such as the driving trajectory of the truck and the停留 time at the entrance and the gantry. Then, combine it with the second vehicle behavior portrait constructed from the radar data, and embed dynamic information such as speed and acceleration into the first vehicle behavior portrait.

[0050] If it is subsequently found that the speed of the truck drops to 40 km / h and lasts for a period of time on a certain section, and at the same time the vehicle distance abnormally shrinks, it can be judged that the truck may have a故障 or abnormal behavior. Mark this behavior in a timely manner, generate the corresponding information board information, and display the information "Truck Ji A12345, during the a-b period, abnormal deceleration, vehicle distance too small, please pay attention to driving norms. If there are special circumstances, you can stop at the emergency lane and contact the rescue if necessary" on the variable message sign that the target vehicle is about to pass, to ensure road safety and normal vehicle driving.

[0051] This embodiment can accurately identify vehicle behavior, and multi-source data complements and verifies each other, such as the combination of checkpoint and ETC gantry video with radar data to fully capture vehicle dynamics, whether it is normal driving, illegal lane change or abnormal speed change, it can accurately locate and reduce misjudgment. This embodiment can also promptly determine vehicle abnormalities and quickly generate information board information, improve the personalization and accuracy of road services, and improve the overall safety and traffic efficiency of highways.

[0052] In one embodiment of the present disclosure, embedding the second vehicle behavior profile into the first vehicle behavior profile based on the driving time sequence and driving position information of the target vehicle includes: A first time-position sequence of the target vehicle is determined based on the first vehicle behavior profile.

[0053] Determine a second time-position sequence of the target vehicle based on the second vehicle behavior portrait. The first time-position sequence and the second time-position sequence each include a plurality of time-position key-value pairs.

[0054] The first time-position sequence and the second time-position sequence are matched based on the target time window and the target sliding step size to obtain a matching path of the first time-position sequence and the second time-position sequence.

[0055] The second vehicle behavior profile is embedded into the first vehicle behavior profile based on a matching path of the first time-position sequence and the second time-position sequence.

[0056] In this embodiment, the first time-position sequence includes a plurality of first key-value pairs extracted from the first vehicle behavior profile, each first key-value pair including a timestamp information and the corresponding position information of the target vehicle. The first time-position sequence records the position sequence of the target vehicle at different times observed through the video data. Each first key-value pair corresponds to a frame of image.

[0057] The second time-position sequence includes multiple second key-value pairs extracted from the second vehicle behavior portrait, each second key-value pair includes a timestamp information and the location information of the target vehicle located by radar, and the second time-position sequence records the location sequence of the target vehicle at different times, and each second key-value pair corresponds to a radar detection data. The target time window is a preset time range parameter, which is used to determine the time range within which the location information corresponding to the two timestamps can be considered to match when matching the first time-position sequence and the second time-position sequence.

[0058] In this embodiment, matching the first time-position sequence and the second time-position sequence based on the target time window and the target sliding step size includes: The first time-position sequence is divided into a plurality of first subsequences based on a target time window and a target sliding step size.

[0059] The second time-position sequence is divided into a plurality of second subsequences based on the target time window and the target sliding step size.

[0060] The plurality of first subsequences and the plurality of second subsequences are nonlinearly aligned in the time dimension to obtain matching paths of the first time-position sequence and the second time-position sequence.

[0061] In this embodiment, non-linearly aligning the plurality of first subsequences and the plurality of second subsequences in the time dimension includes: Calculate target cost values ​​for the first subsequence and the second subsequence based on the cost function.

[0062] Determine the cost matrix based on the target cost value.

[0063] The plurality of first subsequences and the plurality of second subsequences are nonlinearly aligned in a time dimension based on a cost matrix.

[0064] In this embodiment, the cost function is:

[0065] The first subsequence is , represents the i-th element in the first subsequence, that is, the i-th first key-value pair, Represents the timestamp information in the first key-value pair, Represents the position information in the first key-value pair, .

[0066] The second subsequence is , represents the jth element in the second subsequence, that is, the jth second key-value pair, Represents the timestamp information in the second key-value pair, Represents the position information in the second key-value pair, . Represents the weight of the difference between timestamp information, Represents the weight of the difference between position information.

[0067] For example, create a cost matrix C. The dimension of the cost matrix C is (m+1)×(n+1), where m represents the length of the first subsequence and n represents the length of the second subsequence. Initialize C[0][0] to 0, for the first row (j from 1 to n): .here It can be regarded as a virtual starting element, whose timestamp and position can be set according to the specific situation, for example, the timestamp is 0 and the position is (0,0).

[0068] For the first column (i from 1 to m): .here It can be regarded as a virtual starting element, whose timestamp and position can be set according to the specific situation, for example, the timestamp is 0 and the position is (0,0).

[0069] Fill the target cost value calculated by the cost function into the cost matrix.

[0070] Based on the dynamic time warping algorithm, from the lower right corner of the cost matrix Start backtracking to find the minimum cumulative cost path, and use the minimum cumulative cost path as the matching path between the first time-position sequence and the second time-position sequence.

[0071] This embodiment can accurately match the time-position sequence of video and radar data through nonlinear alignment and cost function, integrate information from different sources, and improve the vehicle behavior portrait. This embodiment can handle time expansion and distortion problems and adapt to the situation of inaccurate time synchronization during vehicle driving. The fused behavior portrait contains accurate position, speed, acceleration, etc., which is conducive to subsequent behavior analysis and decision-making, such as traffic violation judgment and road condition assessment, and improves the accuracy and management level of road monitoring.

[0072] In one embodiment of the present disclosure, generating information board information corresponding to a target vehicle based on a target vehicle behavior portrait and road environment monitoring data includes: In response to the road environment monitoring data not exceeding the environment data threshold: marking the target vehicle behavior portrait based on the first rule to obtain target marking information.

[0073] Generate information board information corresponding to the target vehicle based on the target mark information.

[0074] In response to the road environment monitoring data exceeding the environment data threshold: determining a second rule based on the road environment monitoring data exceeding the environment data threshold.

[0075] The target vehicle behavior profile is marked based on the second rule to obtain target marking information.

[0076] The first rule and the second rule are both used to identify and mark vehicle behaviors, and the first rule and the second rule have different criteria for determining vehicle behaviors.

[0077] Generate information board information corresponding to the target vehicle based on the target mark information.

[0078] In this embodiment, the environmental data threshold includes critical values ​​set for various types of road environment monitoring data. The first rule is for normal road environments, and the judgment criteria focus on the vehicle's own normal driving behavior. The second rule is for harsh road environments, focusing on the vehicle's behavior in response to dangerous conditions, such as whether to maintain a safe distance and slow down on slippery roads. The first rule and the second rule have different criteria for judging vehicle behavior and contain different behavior markers.

[0079] Exemplarily, the collected road environment monitoring data is compared with the environmental data threshold. If the threshold is not exceeded, the target vehicle behavior profile is scanned and marked according to the first rule, such as using different colors, numbers, etc. to embed marks in the target vehicle behavior profile, which can represent normal driving or speeding. When the road environment monitoring data exceeds the threshold, according to the second rule in harsh environments, the vehicle's behavior of adapting to poor road conditions is marked, such as not turning on fog lights in foggy days, sudden braking on icy roads, etc., and finally the information board information is generated based on the marks.

[0080] For example, on a main road in a city, the target vehicle is a small car, and its behavior profile records its driving trajectory, speed changes, and whether it changes lanes frequently. Road environment monitoring data covers information such as road wetness, visibility, and wind speed.

[0081] When the road environment monitoring data shows that the road surface is dry, visibility is good, and wind speed is stable, none of which exceeds the environmental data threshold. According to the first rule, it is found that the car has been driving at a constant speed of 50-60 kilometers per hour in the past 10 minutes, without illegal lane changes, and the distance from the car in front is 30 meters (the first rule stipulates that the distance between cars should not be less than 20 meters), marked as "normal driving". The information board information is immediately generated: "License plate number [XX] small car, you are driving steadily, please pay attention to keep a distance."

[0082] A moment later, a heavy rainstorm occurred, the road was slippery, and visibility dropped to 200 meters, exceeding the preset threshold. According to the second rule, it was observed that the car did not slow down, overtook, and was 30 meters away from the car in front (the second rule stipulates that the distance between cars should not be less than 40 meters). It was marked as "Dangerous driving in bad weather, not maintaining a safe distance", and the information board showed: "License plate number [XX] small car, heavy rain, please slow down, maintain a safe distance, and drive carefully", which warned the driver in time to ensure road safety.

[0083] This embodiment can accurately adapt to road conditions. When the environment is good, it urges vehicles to drive in accordance with conventional standards. When the environment is bad, it focuses on dangerous responses to ensure that vehicles always receive appropriate guidance and meet actual needs. This embodiment is real-time and highly targeted, quickly detects vehicle behavior deviations, and provides drivers with targeted reminder services, effectively reducing accident risks. Improve road safety management and optimize traffic order.

[0084] In one embodiment of the present disclosure, in response to identifying that the target vehicle reaches the information board matching position, the information board information is displayed, including: Determine a plurality of vehicle information based on the multi-source monitoring data of the first section. The first section is a preset section range before the information board matching position.

[0085] Determine the information board information display list based on the plurality of vehicle information.

[0086] In response to identifying that the target vehicle reaches the information board matching position, select the information board information corresponding to the target vehicle from the information board information display list for display.

[0087] In this embodiment, the information board information display list includes information board information sequences corresponding to a plurality of vehicles, and each information board information is associated with a specific vehicle.

[0088] In this embodiment, determining the information board information display list based on the plurality of vehicle information includes: Determine a plurality of vehicle information based on the multi-source monitoring data of the first section.

[0089] Determine the information board information corresponding to each vehicle information.

[0090] Determine a display message based on each vehicle information and the information board information corresponding to the vehicle information.

[0091] Determine the information board information display list based on all the display messages.

[0092] Exemplarily, on a highway, there are multiple information boards. One of the information board matching positions is 50 kilometers away from the highway entrance. When a vehicle is driving on the highway, the first section is set to a 2-kilometer range before the information board matching position, that is, the 45 - 50 kilometer section. Within this first section, multi-source monitoring data is collected through devices such as road checkpoint cameras and radars, and the license plate information and appearance information of all vehicles are obtained. Based on this vehicle information, all vehicles and the corresponding information board information of the vehicles are selected to form the information board information display list.

[0093] When the "Guangdong A12345" sedan reaches the information board matching position 50 kilometers away from the entrance, the information board information corresponding to the sedan can be quickly selected from the information board information display list and displayed on the information board to promptly remind the driver to pay attention to the vehicle speed.

[0094] This embodiment generates the information board information display list in advance, enabling quick response. When the vehicle reaches the matching position, relevant information can be quickly displayed.

[0095] In one embodiment of the present disclosure, a vehicle behavior recognition method further includes: The target vehicle's driving route is determined based on the target vehicle's multi-source monitoring data.

[0096] Based on the target vehicle's driving route, multiple historical vehicle behavior portraits and historical intelligence board information corresponding to each historical vehicle behavior portrait are selected from a historical vehicle monitoring database.

[0097] The data matching degree of the multi-source monitoring data of the target vehicle and multiple historical vehicle behavior portraits is calculated to obtain a data matching degree list, and the highest matching degree is determined from the data matching degree list.

[0098] If the highest matching degree is greater than the matching degree threshold, the historical information board information of the historical vehicle behavior portrait corresponding to the highest matching degree is used as the information board information of the target vehicle.

[0099] If the highest matching degree is less than or equal to the matching degree threshold, a behavior profile of the target vehicle is constructed based on the multi-source monitoring data of the target vehicle.

[0100] In this embodiment, the target vehicle driving route is the driving track of the target vehicle on the road. The historical vehicle monitoring database includes a large number of historical vehicle behavior portraits and corresponding information board information. The historical vehicle behavior portrait is a comprehensive description of the behavior characteristics of the historical vehicle during driving. The historical information board information is information generated for the historical vehicle and displayed on the information board. The data matching degree is an indicator to measure the similarity between the multi-source monitoring data of the target vehicle and the historical vehicle behavior portrait. The matching degree threshold is a preset standard value for judging whether the data matching degree is sufficient.

[0101] In this embodiment, when constructing the target vehicle behavior profile based on the road checkpoint video data and the ETC gantry video data, the nearest neighbor algorithm can be used to match the vehicle features. For example, after extracting the key features of the vehicle such as the appearance and license plate, the best matching record is searched in the historical vehicle data set, and its historical trajectory is associated with the current monitoring data to supplement and improve the vehicle's driving path information, making the behavior profile more coherent.

[0102] When analyzing radar detection data, an improved algorithm based on Kalman filtering can be used. Since radar detection data is noisy, the improved algorithm based on Kalman filtering can optimize the estimation of real-time measured speed, acceleration and other information based on the historical motion state of the vehicle. For example, in radar measurements at consecutive time points, through continuous iterative calculations, combined with factors such as vehicle inertia, the estimated values ​​of speed and acceleration are corrected and accurately integrated into the vehicle behavior portrait, improving the accuracy of dynamic driving characteristics.

[0103] An improved algorithm based on the Bayesian filter concept can be used to comprehensively analyze multi-source monitoring data and identify vehicle behavior. A probability model of different vehicle behaviors (such as normal driving, lane changing, speeding, etc.) is established based on historical data. With the input of new multi-source monitoring data, the posterior probability of each behavior is updated, so as to accurately label the vehicle behavior portrait, for example, to determine whether there is an abnormality in the driving of a vehicle on a certain road section.

[0104] In this embodiment, the target vehicle's driving route is determined based on multi-source monitoring data, and the vehicle behavior profile and information board information corresponding to the relevant historical vehicles with the same driving route are screened out from the historical database based on this as a criterion. The data matching degree of the target vehicle and the historical vehicle is calculated, and a matching degree list is obtained and the highest value is found. If the highest matching degree exceeds the preset matching degree threshold, it means that the target vehicle behavior is similar to a certain historical vehicle, and the corresponding historical information board information can be directly used. If it does not exceed the threshold, a new behavior profile is constructed based on the target vehicle data.

[0105] This embodiment can improve the efficiency of information generation by utilizing historical data, and can quickly determine the information on the information board for situations with similar driving routes and behaviors. For situations that do not meet the matching threshold, a new profile is constructed to ensure targeted processing of special situations. The accuracy and efficiency of vehicle behavior recognition are improved, the generation of information on the information board is optimized, and better support is provided for road management and safety assurance.

[0106] In one embodiment of the present disclosure, a vehicle behavior recognition method further includes: The target road is divided based on all the information board matching positions to obtain a plurality of second road sections. There are a plurality of information board matching positions, and each information board matching position corresponds to a second road section.

[0107] If the target vehicle reaches the target information board matching position, multi-source monitoring data is determined based on the second road section corresponding to the target information board matching position.

[0108] In this embodiment, the target road refers to a road where vehicle behavior monitoring and management is being performed.

[0109] By dividing the road into multiple second sections by matching the information board, the target road can be managed in a refined manner, which facilitates targeted analysis of vehicle behavior and avoids repeated calculation of vehicle behavior profiles. When the target vehicle arrives at a specific location, multi-source monitoring data is determined based on the corresponding section, which can accurately obtain detailed information about the vehicle in the section and improve the accuracy of behavior recognition.

[0110] Corresponding to a vehicle behavior recognition method in the above embodiment, Figure 2This is a structural block diagram of a vehicle behavior recognition system provided by an embodiment of the present disclosure. For ease of explanation, only the parts related to the embodiment of the present disclosure are shown. Figure 2 The vehicle behavior recognition system 20 includes: a vehicle behavior portrait module 21, an information board module 22 and a display module 23.

[0111] Among them, the vehicle behavior profiling module 21 is used to construct a target vehicle behavior profile based on multi-source monitoring data of the target vehicle.

[0112] The information board module 22 is used to generate information board information corresponding to the target vehicle based on the target vehicle behavior portrait and road environment monitoring data.

[0113] The display module 23 is used to display the information on the information board in response to identifying that the target vehicle has arrived at the matching position of the information board.

[0114] In one embodiment of the present disclosure, the multi-source monitoring data includes road checkpoint video data, ETC gantry video data and radar detection data. The vehicle behavior profiling module 21 is specifically used to sequentially splice the road checkpoint video data and the ETC gantry video data based on the driving time sequence of the target vehicle to obtain the spliced ​​target video data.

[0115] A first vehicle behavior profile of the target vehicle is determined based on the target video data.

[0116] A second vehicle behavior profile of the target vehicle is determined based on the radar detection data.

[0117] The second vehicle behavior profile is embedded into the first vehicle behavior profile based on the driving time sequence and driving position information of the target vehicle to obtain the target vehicle behavior profile.

[0118] In one embodiment of the present disclosure, the vehicle behavior profiling module 21 is further configured to extract timestamp information and vehicle information of each frame of the target video data.

[0119] Based on the vehicle information of each frame image, the vehicle position information, vehicle appearance features and vehicle behavior information of the target vehicle in the frame image are determined.

[0120] A first vehicle behavior portrait of the target vehicle is constructed based on the timestamp information of each frame image in the target video data, the vehicle position information, the vehicle appearance features and the vehicle behavior information of the target vehicle.

[0121] In one embodiment of the present disclosure, the vehicle behavior profile module 21 is further configured to determine a first time-position sequence of the target vehicle based on the first vehicle behavior profile.

[0122] Determine a second time-position sequence of the target vehicle based on the second vehicle behavior portrait. The first time-position sequence and the second time-position sequence each include a plurality of time-position key-value pairs.

[0123] The first time-position sequence and the second time-position sequence are matched based on the target time window to obtain a matching path of the first time-position sequence and the second time-position sequence.

[0124] The second vehicle behavior profile is embedded into the first vehicle behavior profile based on a matching path of the first time-position sequence and the second time-position sequence.

[0125] In one embodiment of the present disclosure, the information board module 22 is specifically used to, in response to the road environment monitoring data not exceeding the environment data threshold, mark the target vehicle behavior portrait based on the first rule to obtain target marking information.

[0126] Generate information board information corresponding to the target vehicle based on the target mark information.

[0127] In response to the road environment monitoring data exceeding the environment data threshold: determining a second rule based on the road environment monitoring data exceeding the environment data threshold.

[0128] The target vehicle behavior profile is marked based on the second rule to obtain target marking information.

[0129] The first rule and the second rule are both used to identify and mark vehicle behaviors, and the first rule and the second rule have different criteria for determining vehicle behaviors.

[0130] Generate information board information corresponding to the target vehicle based on the target mark information.

[0131] In one embodiment of the present disclosure, the display module 23 is specifically used to determine multiple vehicle information based on multi-source monitoring data of a first road section. The first road section is a preset road section range before the information board matching position.

[0132] An information board information display list is determined based on a plurality of vehicle information.

[0133] In response to identifying that the target vehicle has arrived at the information board matching position, information board information corresponding to the target vehicle is selected from the information board information display list for display.

[0134] In one embodiment of the present disclosure, a vehicle behavior recognition system 20 further includes: The data determination module is used to divide the target road based on all the information board matching positions to obtain multiple second road sections. There are multiple information board matching positions, and each information board matching position corresponds to a second road section.

[0135] If the target vehicle reaches the target information board matching position, multi-source monitoring data is determined based on the second road section corresponding to the target information board matching position.

[0136] See also Figure 3 , Figure 3 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. Figure 3 The electronic device 300 in the embodiment shown may include: one or more processors 301, one or more input devices 302, one or more output devices 303 and one or more memories 304. The processors 301, input devices 302, output devices 303 and memories 304 communicate with each other via a communication bus 305. The memory 304 is used to store computer programs, which include program instructions. The processor 301 is used to execute the program instructions stored in the memory 304. The processor 301 is configured to call the program instructions to execute the functions of the modules in the above-mentioned system embodiments, such as Figure 2 The functions of modules 21 to 23 are shown.

[0137] It should be understood that in the embodiment of the present disclosure, the processor 301 may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0138] The input device 302 may include a touch panel, a fingerprint collection sensor (for collecting the user's fingerprint information and fingerprint direction information), a microphone, etc., and the output device 303 may include a display (LCD, etc.), a speaker, etc.

[0139] The memory 304 may include a read-only memory and a random access memory, and provide instructions and data to the processor 301. A portion of the memory 304 may also include a non-volatile random access memory. For example, the memory 304 may also store information about the device type.

[0140] In a specific implementation, the processor 301, input device 302, and output device 303 described in the embodiments of the present disclosure can execute the implementation methods described in the first and second embodiments of a vehicle behavior recognition method provided in the embodiments of the present disclosure, and can also execute the implementation methods of the electronic device 300 described in the embodiments of the present disclosure, which will not be repeated here.

[0141] In another embodiment of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by the processor, all or part of the processes in the above-mentioned embodiment method are implemented, and the computer program can also be completed by instructing the relevant hardware through the computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0142] The computer-readable storage medium may be an internal storage unit of the electronic device of any of the aforementioned embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Furthermore, the computer-readable storage medium may also include both an internal storage unit of the electronic device and an external storage device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0143] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this disclosure.

[0144] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the electronic devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0145] In the several embodiments provided in the present application, it should be understood that the disclosed electronic device and method can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces or units, or it can be an electrical, mechanical or other form of connection.

[0146] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present disclosure.

[0147] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0148] The above are only specific embodiments of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present disclosure, and these modifications or replacements should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A vehicle behavior recognition method, characterized in that: include: Build a target vehicle behavior profile based on multi-source monitoring data of the target vehicle; Generate information board information corresponding to the target vehicle based on the target vehicle behavior portrait and road environment monitoring data; In response to recognizing that the target vehicle has arrived at the information board matching position, the information board information is displayed.

2. A vehicle behavior recognition method as claimed in claim 1, characterized in that: The multi-source monitoring data includes road checkpoint video data, ETC gantry video data and radar detection data; The target vehicle behavior profile is constructed based on the multi-source monitoring data of the target vehicle, including: Sequentially splicing the road checkpoint video data and the ETC gantry video data based on the driving time sequence of the target vehicle to obtain spliced ​​target video data; Determine a first vehicle behavior portrait of the target vehicle based on the target video data; Determine a second vehicle behavior profile of the target vehicle based on the radar detection data; The second vehicle behavior portrait is embedded into the first vehicle behavior portrait based on the driving time sequence and driving position information of the target vehicle to obtain the target vehicle behavior portrait.

3. A vehicle behavior recognition method as claimed in claim 2, characterized in that: The determining a first vehicle behavior portrait of the target vehicle based on the target video data includes: Extract the timestamp information and vehicle information of each frame image in the target video data; Determine the vehicle position information, vehicle appearance features and vehicle behavior information of the target vehicle in each frame of image based on the vehicle information of the frame of image; A first vehicle behavior portrait of the target vehicle is constructed based on the timestamp information of each frame image in the target video data, the vehicle position information, the vehicle appearance features and the vehicle behavior information of the target vehicle.

4. A vehicle behavior recognition method as claimed in claim 3, characterized in that: The embedding the second vehicle behavior profile into the first vehicle behavior profile based on the driving time sequence and driving position information of the target vehicle includes: Determining a first time-position sequence of the target vehicle based on the first vehicle behavior profile; Determine a second time-position sequence of the target vehicle based on the second vehicle behavior portrait; the first time-position sequence and the second time-position sequence each include a plurality of time-position key-value pairs; Matching the first time-position sequence and the second time-position sequence based on a target time window and a target sliding step size to obtain a matching path of the first time-position sequence and the second time-position sequence; The second vehicle behavior profile is embedded into the first vehicle behavior profile based on a matching path of the first time-position sequence and the second time-position sequence.

5. A vehicle behavior recognition method as claimed in claim 1, characterized in that: The generating of information board information corresponding to the target vehicle based on the target vehicle behavior portrait and road environment monitoring data includes: In response to the road environment monitoring data not exceeding the environment data threshold: marking the target vehicle behavior portrait based on a first rule to obtain target marking information; Generate information board information corresponding to the target vehicle based on the target mark information; In response to the road environment monitoring data exceeding the environment data threshold: determining a second rule based on the road environment monitoring data exceeding the environment data threshold; Marking the target vehicle behavior profile based on the second rule to obtain target marking information; The first rule and the second rule are both used to identify and mark vehicle behavior, and the first rule and the second rule have different criteria for determining vehicle behavior; The information board information corresponding to the target vehicle is generated based on the target mark information.

6. A vehicle behavior recognition method as claimed in claim 1, characterized in that: In response to identifying that the target vehicle has arrived at the information board matching position, displaying the information on the information board includes: Determine multiple vehicle information based on multi-source monitoring data of a first road section; the first road section is a preset road section range before the information board matching position; determining an information display list of the information board based on the plurality of vehicle information; In response to identifying that the target vehicle has arrived at the information board matching position, information board information corresponding to the target vehicle is selected from the information board information display list for display.

7. A vehicle behavior recognition method as claimed in claim 1, characterized in that: Also includes: Dividing the target road based on the matching positions of all information boards to obtain multiple second road sections; There are multiple information board matching positions, and each information board matching position corresponds to a second road section; If the target vehicle reaches the target information board matching position, the multi-source monitoring data is determined based on the second road section corresponding to the target information board matching position.

8. A vehicle behavior recognition system, characterized in that: include: The vehicle behavior profiling module is used to construct a target vehicle behavior profile based on the multi-source monitoring data of the target vehicle; An information board module, used to generate information board information corresponding to the target vehicle based on the target vehicle behavior portrait and road environment monitoring data; The display module is used to display the information on the information board in response to identifying that the target vehicle has arrived at the matching position of the information board.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing 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 7 are implemented.

Images