An intelligent traffic control method and device based on AI visual analysis and a medium

Abstract

The application discloses a kind of intelligent traffic control method, device and medium based on AI vision analysis, including for controlling right turn vehicle and electric vehicle grouping traffic, it is characterized in that, including the following steps: collecting right turn lane and non-motor vehicle lane video;Vehicle information and electric vehicle information are obtained by analyzing video;Whether grouping traffic is judged according to the information, the spatial clustering of electric vehicle in second information is carried out using DBSCAN density clustering algorithm, and preliminary clustering result is formed;Multi-dimensional grouping parameter model is constructed, and dynamic weight is assigned to each parameter;Based on the dynamic weight and preliminary clustering result, electric vehicle traffic grouping is generated by multi-objective optimization algorithm;According to the electric vehicle traffic grouping, control electric vehicle to cross right turn lane in turn as unit group.The application can greatly improve electric vehicle and right turn vehicle traffic efficiency and safety.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of AI visual analysis, in particular to a smart traffic control method and device based on AI visual analysis and a medium. BACKGROUND

[0002] In the field of road transportation, national highways serve as important traffic trunks and bear a large amount of passenger and freight transportation tasks. At present, many right-turn national highways have significant deficiencies in traffic facility configuration, and a dedicated right-turn traffic signal is generally not set. At the same time, the waiting areas for pedestrians, electric vehicles and bicycles are mostly planned on the left side of the right-turn lane, which causes the pedestrians, electric vehicles and bicycles to cross the driving path of the right-turn vehicles when entering the waiting area.

[0003] In this process, since the right-turn vehicles and the pedestrians, electric vehicles and bicycles lack corresponding traffic signal indications, the vehicles and the pedestrians, electric vehicles and bicycles frequently compete for passage. Such disordered passage not only greatly increases the risk of traffic accidents, seriously threatening the life safety of road users, but also causes traffic congestion, significantly reduces the road passage efficiency and destroys the normal traffic order.

[0004] In view of the above problems, the existing technology mainly relies on traffic management personnel to conduct on-site guidance. However, this method has obvious drawbacks, which not only requires a large amount of human cost, but also has low efficiency in the face of complex and variable traffic conditions, and it is difficult to fundamentally solve the problem of vehicle-pedestrian conflict. Therefore, it is urgent to develop a new traffic control technology to overcome the shortcomings of the existing technology, ensure traffic safety and improve road passage efficiency. SUMMARY

[0005] In view of the above problems, the present application provides a smart traffic control method and device based on AI visual analysis to solve the technical problems of right-turn vehicle and electric vehicle passage confusion and ineffective guidance.

[0006] To achieve the above purpose, the present application provides a smart traffic control method based on AI visual analysis for controlling the grouped passage of right-turn vehicles and electric vehicles, comprising the following steps:

[0007] Step 1: video acquisition, acquiring a first video of the right-turn lane through a first camera and acquiring a second video of the non-motor vehicle lane on the left side of the right-turn vehicle through a second camera;

[0008] Step 2: AI video analysis, comprising:

[0009] 2.1 preprocessing the second video to remove useless information and extract effective features to obtain the position coordinates, motion state, waiting time and spatial distribution of the electric vehicle as basic data;

[0010] 2.2 Analyzing the first video to obtain first information of the number of right-turn vehicles, vehicle speed, waiting time length, and time period characteristics; further analyzing the preprocessed second video data to obtain second information of the number of electric vehicles, density, gathering area, and waiting time length;

[0011] Judging whether to perform step 3 based on the first information and the second information; the judging standard is that the number of right-turn vehicles is greater than or equal to a first threshold value, or the number of electric vehicles is greater than a second threshold value;

[0012] Step 3: Intelligent grouping, including:

[0013] 3.1 Using a DBSCAN density clustering algorithm to perform spatial clustering on the electric vehicles in the second information to form a preliminary clustering result;

[0014] 3.2 Constructing a multi-dimensional grouping parameter model, the parameters including the traffic volume, average speed, average waiting time length, and current time period characteristics of the right-turn vehicles, and assigning dynamic weights to each parameter;

[0015] 3.3 Generating electric vehicle passing groups based on the dynamic weights and the preliminary clustering result through a multi-objective optimization algorithm;

[0016] Step 4: Passing control, controlling the electric vehicles to pass through the right-turn lane in groups in sequence according to the electric vehicle passing groups, and informing the right-turn vehicles to stop and wait while the electric vehicles are passing.

[0017] To solve the above technical problems, another technical solution is provided:

[0018] An intelligent traffic control device based on AI visual analysis, comprising:

[0019] A video acquisition module acquires a first video of a right-turn lane through a first camera and acquires a second video of a non-motor vehicle lane on the left side of a right-turn vehicle through a second camera;

[0020] An AI video analysis module is configured to preprocess the second video to remove useless information and extract effective features, to obtain basic data of the position coordinates, motion state, waiting time length, and spatial distribution of electric vehicles; to analyze the first video to obtain first information of the number of right-turn vehicles, vehicle speed, waiting time length, and time period characteristics; and to further analyze the preprocessed second video data to obtain second information of the number of electric vehicles, density, gathering area, and waiting time length;

[0021] A judging module is configured to judge whether to perform step 3 based on the first information and the second information; the judging standard is that the number of right-turn vehicles is greater than or equal to a first threshold value, or the number of electric vehicles is greater than a second threshold value;

[0022] The intelligent grouping module is configured to perform spatial clustering on the electric vehicles in the second information by using a DBSCAN density clustering algorithm to form a preliminary clustering result; a multi-dimensional grouping parameter model is constructed, the parameters including the traffic volume, average speed, average waiting time and current time period characteristics of the right-turn vehicles, and dynamic weights are assigned to each parameter; and the electric vehicle traffic grouping is generated by using a multi-objective optimization algorithm based on the dynamic weights and the preliminary clustering result;

[0023] The traffic control module is configured to control the electric vehicles to pass through the right-turn lane in groups according to the electric vehicle traffic grouping, and to inform the right-turn vehicles to stop and wait while the electric vehicles are passing.

[0024] The application also provides a computer readable storage medium having a computer program stored therein, wherein the computer program, when executed by a processor, performs the AI vision analysis-based intelligent traffic control method described in any one of the technical solutions above.

[0025] Unlike the prior art, the above technical solution can obtain the traffic information of the right-turn vehicles and the electric vehicles in real time and accurately by using double cameras to collect videos and combining AI vision analysis, and can provide data support for traffic control; and in the present application, the DBSCAN clustering algorithm is used to group the electric vehicles in space, and the grouping parameters are dynamically adjusted according to the real-time state of the vehicles, so that the intelligent grouping of the electric vehicles is realized, which takes into account the aggregation characteristics of the electric vehicles and also takes into account the waiting fairness of the vehicles. The electric vehicle traffic is controlled by grouping, and the vehicles are informed to wait at the same time, effectively solving the traffic conflict between the right-turn vehicles and the electric vehicles, improving the intersection traffic efficiency, and reducing the risk of traffic accidents.

[0026] The above summary of the invention is only a summary of the technical solutions of the present application, in order to enable those skilled in the art to more clearly understand the technical solutions of the present application, and then to implement the content recorded in the specification and drawings, and in order to make the above and other purposes, characteristics and advantages of the present application more easily understood, the following will be described in combination with the specific embodiments of the present application and the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0027] The accompanying drawings are only used to illustrate the principles, implementation manners, applications, characteristics and effects of the specific embodiments and other related contents of the present application, and cannot be considered as limitations of the present application.

[0028] In the drawings of the specification:

[0029] Figure 1 The flowchart of the AI vision analysis-based intelligent traffic control method described in the specific embodiments;

[0030] Figure 2Flow chart of the intelligent grouping described for the specific embodiments

[0031] Figure 3 Flow chart of the pre-processing of the second video described for the specific embodiments

[0032] Figure 4 Flow chart of the construction of the multi-dimensional grouping parameter model described for the specific embodiments

[0033] Figure 5 Module block diagram of the intelligent traffic control device based on AI visual analysis described for the specific embodiments

[0034] Figure 6 Schematic diagram of the computer readable storage medium described for the specific embodiments

[0035] The reference signs involved in the above-mentioned various figures are explained as follows:

[0036] 500, intelligent traffic control device based on AI visual analysis; 501, video acquisition module; 502, AI video analysis module; 503, judgment module; 504, intelligent grouping module; 505, traffic control module

[0037] 600, computer readable storage medium DETAILED DESCRIPTION

[0038] In order to describe the possible application scenarios, technical principles, specific schemes that can be implemented, purposes and effects that can be achieved, etc. of the present application in detail, the following will be described in detail in combination with the specific embodiments listed and with the aid of the drawings. The embodiments described in this paper are only used to more clearly illustrate the technical solutions of the present application, and therefore only serve as examples, and cannot limit the protection scope of the present application.

[0039] In this paper, the term "embodiment" means that the specific features, structures or characteristics described in combination with the embodiment can be included in at least one embodiment of the present application. The term "embodiment" appearing at various places in the specification does not necessarily refer to the same embodiment, and does not particularly limit the independence or association between other embodiments. In principle, in this application, as long as there is no technical contradiction or conflict, the technical features mentioned in each embodiment can be combined in any way to form a corresponding implementable technical solution.

[0040] Unless otherwise defined, the meanings of the technical terms used in this paper are the same as those commonly understood by the person skilled in the art to which the present application belongs; the use of related terms in this paper is only for the purpose of describing specific embodiments, and is not intended to limit the present application.

[0041] In the description of the present application, the phrase "and / or" is a description of a logical relationship between objects, which means that there can be three relationships, for example, A and / or B, which means that there are three cases: A exists, B exists, and A and B exist at the same time. In addition, the character " / " herein generally represents that the associated objects before and after are a "or" logical relationship.

[0042] In the present application, the terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual quantity, primary and secondary or order relationship between the entities or operations.

[0043] In the present application, without more limitation, the "include", "contain", "have" or other similar open expressions used in the sentence are intended to cover non-exclusive inclusion, and these expressions do not exclude the existence of other elements in the process, method or product including the described elements, so that the process, method or product including a series of elements can not only include those limited elements, but also include other elements not explicitly listed, or also include the elements inherent in such process, method or product.

[0044] As the same understanding as in the "Guidelines for Examination", in the present application, the expressions such as "greater than", "less than", "exceed" are understood as not including the number; the expressions such as "above", "below", "within" are understood as including the number. In addition, in the description of the embodiments of the present application, the meaning of "multiple" is more than two (including two), and similar expressions related to "multiple" are also understood in this way, for example, "multiple groups", "multiple times" and the like, unless otherwise explicitly limited.

[0045] In the description of the embodiments of the present application, the spatial-related expressions used, such as "center", "longitudinal", "transverse", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "vertical", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", and the like, indicate the orientation or positional relationship based on the orientation or positional relationship shown in the specific embodiments or the drawings, and are only for the convenience of describing the specific embodiments of the present application or for the reader to understand, and do not indicate or imply that the indicated device or component must have a specific position, a specific orientation, or be constructed or operated in a specific orientation, and therefore cannot be understood as a limitation on the embodiments of the present application.

[0046] Unless otherwise clearly indicated or limited by context, reference in the specification to "installation", "connection", "coupling", "fixing", "setting", and the like, is to be construed as being broad in nature, for example "connection" can be a fixed connection, a detachable connection, or an integral setting; it can be a mechanical connection, an electrical connection, or a communication connection; it can be a direct connection, or an indirect connection via an intermediate medium; it can be an internal connection of two elements, or an interaction relationship between two elements. For those skilled in the art to which the present application belongs, the specific meanings of the above-mentioned terms in the embodiments of the present application can be understood according to the specific circumstances.

[0047] Please refer to Figures 1 to 4 The embodiment provides a smart traffic control method based on AI visual analysis. The smart traffic control method based on AI visual analysis is applied to a cross intersection of a city road without right-turn red light, and is used for controlling right-turn vehicles and electric vehicle grouping to pass through, so as to improve the safety and passing efficiency of the electric vehicle passing through the right-turn lane. The electric vehicle refers to an electric bicycle with two wheels, rather than an electric car with four wheels. As shown in the figure, Figure 1 The smart traffic control method based on AI visual analysis comprises the following steps:

[0048] Step 1: video acquisition, the first video of the right-turn lane is acquired by a first camera, and the second video of the left non-motor vehicle lane of the right-turn vehicle is acquired by a second camera. The first camera can be installed 5 meters in front of the right-turn lane stop line, the model of the first camera is Hikvision DS-2CD3T25-I5 (2000 pixels, focal length 4mm, frame rate 25fps), and the first camera covers the right-turn lane 100 meters and the waiting area. The second camera is installed on the east sidewalk lamp pole (height 3.5 meters), the model is the same as above, the lens is directed to the northwest direction, and the second camera covers the sidewalk area of 20m*5m. The edge computing device is NVIDIA Jetson AGX Xavier (8-core ARM CPU, 512-core GPU), which is responsible for video processing and algorithm running.

[0049] Step 2: video analysis, comprising:

[0050] 2.1, the second video is preprocessed to remove useless information and extract effective features, and the position coordinates, motion state, waiting time and spatial distribution of the electric vehicle are obtained. Basic data.

[0051] 2.2 Analyze the first video to obtain first information of the number of right-turn vehicles, vehicle speed, waiting time, and time period characteristics; further analyze the pre-processed second video data to obtain second information of the number of electric vehicles, density, gathering area, and waiting time; determine whether to perform step 3 based on the first information and the second information; the criteria for the determination are that the number of right-turn vehicles is greater than or equal to a first threshold value (such as 10 vehicles), or the number of electric vehicles is greater than a second threshold value (such as 20 vehicles). In step 2.2, the analysis of the first video uses the YOLOv8 model to detect right-turn vehicles, and the DeepSORT algorithm is used for continuous tracking of the vehicles, and the actual vehicle speed is calculated by the inter-frame displacement and the camera calibration parameters.

[0052] Step 3: Intelligent grouping, as shown in Figure 2 , which includes:

[0053] 3.1 Using the DBSCAN density clustering algorithm to perform spatial clustering on the electric vehicles in the second information to form a preliminary clustering result.

[0054] 3.2 Constructing a multi-dimensional grouping parameter model, the parameters including the traffic volume, average speed, average waiting time, and current time period characteristics of the right-turn vehicles, and assigning dynamic weights to each parameter.

[0055] 3.3 Based on the dynamic weights and the preliminary clustering result, generating an electric vehicle passing grouping through a multi-objective optimization algorithm.

[0056] Step 4: Passing control, according to the electric vehicle passing grouping, controlling the electric vehicles to pass through the right-turn lane in groups, while informing the right-turn vehicles to stop and wait when the electric vehicles are passing.

[0057] In step 3.1, the pixel coordinates of the electric vehicles are converted to actual coordinates (with the camera as the origin, the X-axis along the non-motor vehicle lane direction, and the Y-axis perpendicular to the X-axis), the ε of the DBSCAN algorithm is set to 2.0 meters, and the min_samples is set to 3. The electric vehicles in the waiting area are clustered to obtain 3 initial clustering clusters (containing 6, 8, and 5 electric vehicles respectively). The basic grouping parameters are set as follows: basic group size = 6 vehicles, maximum number of groups = 3 groups, and group passing interval = 4-10 seconds.

[0058] The grouping result of the intelligent grouping can be projected on the non-motor vehicle lane by an infrared projector, for example, by projecting rectangular frames to frame the 1st group and the 2nd group respectively, so that the electric vehicle riders know their grouping result and perform grouping passing according to the prompt.

[0059] During traffic control, when the first group of electric vehicles passes, the right-turn lane LED screen displays "Electric vehicles passing, please wait" and the red light illuminates. The non-motorized vehicle lane LED screen displays "First group of electric vehicles passing" for 20 seconds (calculated based on the group size and the average speed of the electric vehicles). After the first group completes its passage, the second group begins to pass after a 4-second interval, and the above display control repeats. After all three groups of electric vehicles have passed, the right-turn lane LED screen displays "Vehicles can pass," the red light turns off, and the green light illuminates, allowing vehicles to turn right.

[0060] In this embodiment, a closed-loop control system is implemented through four core steps: video capture, video analysis, intelligent grouping, and traffic control. Specifically, dual cameras are used for video capture, ensuring simultaneous coverage of right-turning vehicles and the area occupied by electric vehicles in non-motorized lanes. A targeted preprocessing step is added to the video analysis phase to address the issues of target clutter and background interference encountered in traditional methods. Intelligent grouping incorporates a multidimensional parameter model and optimization algorithm to overcome the limitations of single-factor decision-making. Traffic control achieves coordinated coordination between vehicles and electric vehicles through two-way information exchange.

[0061] This embodiment enables active coordinated control of right-turning vehicles and electric vehicles at intersections without traffic lights, fundamentally resolving traffic conflicts under traditional traffic rules and effectively reducing traffic jams at intersections. This embodiment establishes a complete AI visual analysis closed loop, with an end-to-end processing latency of ≤500ms from raw video to control commands, meeting real-time traffic control requirements.

[0062] In this embodiment, in order to obtain the second information of the electric vehicle more efficiently and accurately, such as Figure 3 As shown, the pre-processing of the second video in step 2.1 includes:

[0063] 2.1.1 Target filtering: The target detection model is used to identify the electric vehicle in the second video and filter out useless targets such as pedestrians, non-motor vehicles, and stationary obstacles.

[0064] 2.1.2 Area division: The video screen is divided into waiting area, transition area and passage area through camera calibration, and only the electric vehicle data in the waiting area is retained.

[0065] 2.1.3 Feature extraction: Extract the center coordinates, motion trajectory, waiting time, and velocity vector features of the electric vehicle, and remove temporary targets with chaotic trajectories or stay times less than 3 seconds.

[0066] The target detection model in step 2.1.1 is a YOLOv8 model fine-tuned based on an electric vehicle sample set. The sample set contains images of electric vehicles under different lighting and weather conditions, and the outline and riding status of the electric vehicles are marked.

[0067] In this embodiment, the demand for motor vehicle detection is achieved by YOLOv8 to realize fast and accurate vehicle positioning, combined with DeepSORT algorithm to solve the problem of vehicle occlusion and cross-frame tracking, and the conversion from pixels to actual physical quantities is realized through camera calibration. In this embodiment, in view of the particularity of electric vehicle detection, the model's recognition ability for electric vehicles is enhanced through special sample set training (covering complex scenes such as backlight and rain) and feature fine-tuning, especially the ability to distinguish electric vehicles from other non-motor vehicles.

[0068] As shown in Figure 4 , the step 3.2 of constructing a multi-dimensional grouping parameter model includes:

[0069] 3.2.1 Determine the core parameter set: including vehicle dimension parameters (vehicle flow Q, average vehicle speed V, and average waiting time T) and time dimension parameters (peak coefficient S, peak period S=1.2, flat peak period S=1.0, and low peak period S=0.8).

[0070] 3.2.2 Dynamic weight distribution: the analytic hierarchy process (AHP) is used to determine the weight of each parameter, wherein the weight of vehicle flow ω1=0.3, the weight of average vehicle speed ω2=0.25, the weight of average waiting time ω3=0.3, and the weight of peak coefficient ω4=0.15.

[0071] 3.2.3 Comprehensive score calculation: the comprehensive score of vehicle state is calculated by the formula C=ω1×Q+ω2×(1 / V)+ω3×T+ω4×S, wherein the higher the value of C, the greater the waiting pressure of the vehicle.

[0072] The step 3.3 of generating electric vehicle passing grouping by a multi-objective optimization algorithm specifically includes:

[0073] 3.3.1 Set optimization goal: minimize total waiting time of electric vehicles and minimize average waiting time of vehicles.

[0074] 3.3.2 Dynamically adjust grouping parameters based on comprehensive score C: when C≥80, the upper limit of group size=10 vehicles, the maximum number of groups=2 groups, and the interval between groups=4 seconds; when 50≤C<80, the upper limit of group size=15 vehicles, the maximum number of groups=3 groups, and the interval between groups=7 seconds; when C<50, the upper limit of group size=20 vehicles, the maximum number of groups=5 groups, and the interval between groups=10 seconds.

[0075] 3.3.3 Optimize the preliminary clustering results by using genetic algorithm, and preferentially allocate the electric vehicles with the longest waiting time to the early passing group, while ensuring that the spatial aggregation degree of each group is ≥0.8 (aggregation degree=average distance of electric vehicles in the group / average distance between groups).

[0076] In this embodiment, the video data processing and intelligent grouping mechanism are optimized, so that the grouping of electric vehicles can be more reasonable and accurate. In this embodiment, the multi-dimensional parameter model breaks through the limitations of traditional single factor decision-making, quantifies the traffic flow, vehicle speed, waiting time and time period into a calculable comprehensive score, and realizes the accurate description of the traffic state. And the multi-objective optimization algorithm (genetic algorithm) realizes the triple balance of "efficiency-fairness-safety" through dynamic parameter adjustment and priority sorting on the basis of spatial clustering.

[0077] The fitness function of the genetic algorithm is F = a x (1 / total waiting time of electric vehicles) + b x (1 / average waiting time of vehicles), wherein a and b are weight coefficients, and a + b = 1, and the peak coefficient S is dynamically adjusted (a = 0.4, b = 0.6 during peak period; a = 0.5, b = 0.5 during flat peak period).

[0078] In this embodiment, the core component of the genetic algorithm, the fitness function, is further optimized, and the dynamic weight coefficients a and b are used to achieve differentiated optimization goals in different traffic periods: the vehicle passing weight is increased (b = 0.6) during the peak period (S = 1.2) to prioritize the vehicle passing efficiency of the main road; and the balanced weight (a = 0.5, b = 0.5) is used during the flat peak period (S = 1.0) to balance the passing rights and interests of vehicles and electric vehicles.

[0079] As shown in Figure 5 In one embodiment, an AI vision analysis-based intelligent traffic control device 500 is provided. The AI vision analysis-based intelligent traffic control device is used to control the grouping of right-turn vehicles and electric vehicles to improve the safety and efficiency of electric vehicle passing. The AI vision analysis-based intelligent traffic control device 500 includes a video acquisition module 501, an AI video analysis module 502, a judgment module 503, an intelligent grouping module 504, and a passing control module 505.

[0080] The video acquisition module 501 acquires a first video of a right-turn lane through a first camera and acquires a second video of a non-motor vehicle lane on the left side of a right-turn vehicle through a second camera.

[0081] The AI video analysis module 502 is used to preprocess the second video, remove useless information, and extract effective features to obtain the position coordinates, motion state, waiting time, and spatial distribution of the electric vehicle; analyze the first video to obtain the number, speed, waiting time, and time period characteristics of the right-turn vehicles; and further analyze the preprocessed second video data to obtain the number, density, aggregation area, and waiting time of the electric vehicles.

[0082] The judgment module 503 is used to judge whether to execute step 3 based on the first information and the second information; the judgment standard is that the number of right-turning vehicles is greater than or equal to a first threshold, or the number of electric vehicles is greater than a second threshold.

[0083] The intelligent grouping module 504 is used to spatially cluster the electric vehicles in the second information using the DBSCAN density clustering algorithm to form a preliminary clustering result; construct a multidimensional grouping parameter model, wherein the parameters include the traffic flow, average speed, average waiting time and current time period characteristics of right-turning vehicles, and assign dynamic weights to each parameter; based on the dynamic weights and the preliminary clustering results, generate electric vehicle traffic groups through a multi-objective optimization algorithm.

[0084] The traffic control module 505 controls the electric vehicles to pass through the right-turn lane in groups according to the traffic grouping of the electric vehicles, and at the same time notifies right-turning vehicles to stop and wait when the electric vehicles pass.

[0085] In this embodiment, the YOLOv8 model is used to analyze the first video to detect right-turning vehicles, the DeepSORT algorithm is used to continuously track the vehicle, and the actual vehicle speed is calculated through the inter-frame displacement and camera calibration parameters.

[0086] like Figure 6 As shown, in another embodiment, a computer-readable storage medium 600 is provided, which stores a computer program. When the computer program is run by a processor, it executes the intelligent traffic control method based on AI visual analysis described in the above embodiment.

[0087] The computer-readable storage medium may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. The non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a ferromagnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM); the magnetic surface memory may be a magnetic disk memory or a magnetic tape memory.

[0088] The volatile memory can be a Random Access Memory (RAM) used as an external cache. By way of example, and not limitation, many forms of RAM can be used, such as Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDR SDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), Sync Link Dynamic Random Access Memory (SLDRAM), Direct Rambus Random Access Memory (DRRAM). The computer readable storage medium described in the embodiment of the present invention is intended to include these and any other suitable types of memory.

[0089] In some embodiments, the processor can be implemented by software, hardware, firmware or a combination thereof, and can use at least one of circuit, single or multiple Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), controller, microcontroller, microprocessor, so that the processor can execute part or all of the steps or any combination of the steps in the intelligent traffic control method based on AI vision analysis in various embodiments of the present application.

[0090] Finally, it should be noted that although the above embodiments have been described in the specification and drawings of the present application, the patent protection scope of the present application should not be limited thereby. Any technical solutions obtained by replacing or modifying the equivalent structure or equivalent flow based on the essential concept of the present application, using the content described in the specification and drawings of the present application, and directly or indirectly implementing the technical solutions of the above embodiments in other related technical fields, etc., are all included in the patent protection scope of the present application.

Claims

1. A smart traffic control method based on AI visual analysis, used to control the grouping of right-turning vehicles and electric vehicles, characterized in that: The following steps are involved: Step 1: Video acquisition: a first camera captures a first video of the right-turn lane, and a second camera captures a second video of the non-motorized vehicle lane on the left side of the right-turn vehicle; Step 2: AI video analysis, including: 2.1 Preprocess the second video to remove useless information and extract effective features to obtain basic data on the electric vehicle's location coordinates, motion status, waiting time, and spatial distribution; 2.2 Analyze the first video to obtain first information on the number, speed, waiting time, and time period characteristics of right-turning vehicles; further analyze the pre-processed second video data to obtain second information on the number, density, gathering area, and waiting time of electric vehicles; Determining whether to execute step 3 based on the first information and the second information; the criterion for the determination is that the number of right-turning vehicles is greater than or equal to a first threshold, or the number of electric vehicles is greater than a second threshold; Step 3: Intelligent grouping, including: 3.1 Use the DBSCAN density clustering algorithm to spatially cluster the electric vehicles in the second information to form a preliminary clustering result; 3.2 Construct a multi-dimensional grouping parameter model, where the parameters include right-turn vehicle flow, average speed, average waiting time, and current time period characteristics, and assign dynamic weights to each parameter; 3.3 Based on the dynamic weights and preliminary clustering results, electric vehicle traffic groups are generated through a multi-objective optimization algorithm; Step 4: Traffic control, according to the electric vehicle traffic grouping, controlling the electric vehicles to pass through the right turn lane in groups, and notifying right-turning vehicles to stop and wait when the electric vehicles pass; The step 3.2 of constructing the multidimensional grouping parameter model includes: 3.2.1 Determine the core parameter set: including vehicle dimension parameters and time dimension parameters. The vehicle dimension parameters include traffic flow Q, average speed V, and average waiting time T; 3.2.2 Dynamic Weight Allocation: The analytic hierarchy process is used to determine the weights of each parameter, where the weight of traffic flow is ω1 = 0.3, the weight of average speed is ω2 = 0.25, the weight of average waiting time is ω3 = 0.3, and the weight of peak coefficient is ω4 = 0.15; 3.2.3 Comprehensive Score Calculation: The comprehensive vehicle status score is calculated using the formula C = ω1 × Q + ω2 × (1 / V) + ω3 × T + ω4 × S. A higher C value indicates greater waiting pressure for the vehicle. Q represents the traffic volume, V represents the average speed, T represents the average waiting time, and S represents the time dimension parameter. During peak hours, S = 1.2, during off-peak hours, S = 1.0, and during low-peak hours, S = 0.

8. The generation of electric vehicle traffic groups by the multi-objective optimization algorithm in step 3.3 specifically includes: 3.3.1 Set optimization goals: minimize the total waiting time of electric vehicles and minimize the average waiting time of vehicles; 3.3.2 Dynamically adjust grouping parameters based on the comprehensive score C: When C ≥ 80, the maximum group size is 10 vehicles, the maximum number of groups is 2, and the interval between groups is 4 seconds; when 50 ≤ C < 80, the maximum group size is 15 vehicles, the maximum number of groups is 3, and the interval between groups is 7 seconds; when C < 50, the maximum group size is 20 vehicles, the maximum number of groups is 5, and the interval between groups is 10 seconds; 3.3.3 A genetic algorithm is used to optimize the preliminary clustering results, prioritizing the allocation of electric vehicles with the longest waiting times to the early passage group, while ensuring that the spatial aggregation degree of each group is ≥ 0.8; aggregation degree = average distance between electric vehicles within a group / average distance between groups.

2. The intelligent traffic control method based on AI visual analysis according to claim 1, wherein the pre-processing of the second video in step 2.1 comprises: 2.1.1 Target Filtering: Use the target detection model to identify the electric vehicle in the second video and filter out useless targets such as pedestrians, non-motor vehicles, and stationary obstacles; 2.1.2 Area Division: The video screen is divided into waiting area, transition area and passage area through camera calibration, and only the electric vehicle data in the waiting area is retained; 2.1.3 Feature extraction: Extract the center coordinates, motion trajectory, waiting time, and velocity vector features of the electric vehicle, and remove temporary targets with chaotic trajectories or stay times less than 5 seconds.

3. The intelligent traffic control method based on AI visual analysis according to claim 1 is characterized in that: The analysis of the first video in step 2.2 uses the YOLOv8 model to detect right-turning vehicles, uses the DeepSORT algorithm to continuously track the vehicles, and calculates the actual vehicle speed through the inter-frame displacement and camera calibration parameters.

4. The intelligent traffic control method based on AI visual analysis according to claim 1 is characterized in that: The target detection model in step 2.1.1 is a YOLOv8 model fine-tuned based on an electric vehicle sample set. The sample set contains images of electric vehicles under different lighting and weather conditions, and the outline and riding status of the electric vehicles are marked.

5. The intelligent traffic control method based on AI visual analysis according to claim 1 is characterized in that: The fitness function of the genetic algorithm is F=α×(1 / total waiting time of electric vehicles)+β×(1 / average waiting time of vehicles), where α and β are weight coefficients, and α+β=1. It is dynamically adjusted according to the peak coefficient S. During peak hours, α=0.4 and β=0.6; during off-peak hours, α=0.5 and β=0.

5.

6. An intelligent traffic control device based on AI visual analysis, characterized in that: include: A video acquisition module, which acquires a first video of the right-turn lane through a first camera, and acquires a second video of the non-motorized vehicle lane on the left side of the right-turn vehicle through a second camera; The AI ​​video analysis module is used to pre-process the second video, remove useless information and extract effective features to obtain basic data on the electric vehicle's position coordinates, motion status, waiting time and spatial distribution; analyze the first video to obtain first information on the number of right-turning vehicles, vehicle speed, waiting time and time period characteristics; Further analyzing the pre-processed second video data to obtain second information on the number, density, gathering area, and waiting time of electric vehicles; a judgment module, configured to judge whether to execute step 3 based on the first information and the second information; the judgment criterion being that the number of right-turning vehicles is greater than or equal to a first threshold, or the number of electric vehicles is greater than a second threshold; an intelligent grouping module for spatially clustering the electric vehicles in the second information using a DBSCAN density clustering algorithm to form a preliminary clustering result; constructing a multidimensional grouping parameter model, wherein the parameters include the traffic volume, average speed, average waiting time, and current time period characteristics of right-turning vehicles, and assigning a dynamic weight to each parameter; Based on the dynamic weights and preliminary clustering results, generating electric vehicle traffic groups through a multi-objective optimization algorithm; A traffic control module controls the electric vehicles to pass through the right-turn lane in groups according to the electric vehicle traffic grouping, and notifies right-turning vehicles to stop and wait when the electric vehicles pass; Constructing a multidimensional grouping parameter model includes: 3.2.1 Determine the core parameter set: including vehicle dimension parameters and time dimension parameters. The vehicle dimension parameters include traffic flow Q, average speed V, and average waiting time T; 3.2.2 Dynamic Weight Allocation: The analytic hierarchy process is used to determine the weights of each parameter, where the weight of traffic flow is ω1 = 0.3, the weight of average speed is ω2 = 0.25, the weight of average waiting time is ω3 = 0.3, and the weight of peak coefficient is ω4 = 0.15; 3.2.3 Comprehensive Score Calculation: The comprehensive vehicle status score is calculated using the formula C = ω1 × Q + ω2 × (1 / V) + ω3 × T + ω4 × S. A higher C value indicates greater waiting pressure for the vehicle. Q represents the traffic volume, V represents the average speed, T represents the average waiting time, and S represents the time dimension parameter. During peak hours, S = 1.2, during off-peak hours, S = 1.0, and during low-peak hours, S = 0.

8. The generation of electric vehicle traffic groups by the multi-objective optimization algorithm in step 3.3 specifically includes: 3.3.1 Set optimization goals: minimize the total waiting time of electric vehicles and minimize the average waiting time of vehicles; 3.3.2 Dynamically adjust grouping parameters based on the comprehensive score C: When C ≥ 80, the maximum group size is 10 vehicles, the maximum number of groups is 2, and the interval between groups is 4 seconds; when 50 ≤ C < 80, the maximum group size is 15 vehicles, the maximum number of groups is 3, and the interval between groups is 7 seconds; when C < 50, the maximum group size is 20 vehicles, the maximum number of groups is 5, and the interval between groups is 10 seconds; 3.3.3 A genetic algorithm is used to optimize the preliminary clustering results, prioritizing the allocation of electric vehicles with the longest waiting times to the early passage group, while ensuring that the spatial aggregation degree of each group is ≥ 0.8; aggregation degree = average distance between electric vehicles within a group / average distance between groups.

7. The intelligent traffic control device based on AI visual analysis according to claim 6 is characterized in that: The analysis of the first video uses the YOLOv8 model to detect right-turning vehicles, the DeepSORT algorithm to continuously track the vehicles, and the actual vehicle speed is calculated through the inter-frame displacement and camera calibration parameters.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, the intelligent traffic control method based on AI visual analysis according to any one of claims 1 to 5 is executed.

Images