Highway tunnel active warning method and system based on dangerous driving behavior

Abstract

The application discloses a highway tunnel active early warning method and system based on dangerous driving behaviors, and the system comprises an acquisition module and a projection module. First, the acquisition module acquires the driving speed of a dangerous driving vehicle, the reading time of a driver to a warning sign, the visual half angle of a driver's visual conical angle, and the height between the human eye and the highest point of effective information of the warning sign. Then, the projection module calculates the shortest visible distance of the driver to the warning sign according to the formation speed, the reading time, the visual half angle and the height. After the shortest visible distance is calculated, the projection module can determine the visible distance range of the warning sign according to the shortest visible distance and the projection position of the warning sign. When the dangerous driving vehicle travels to a position with a distance greater than the shortest visible distance from the projection position of the warning sign, the projection device can project the warning sign at the projection position.

Description

Technical Field

[0001] This invention relates to the field of traffic control technology for road vehicles, and specifically to a method and system for proactive early warning of dangerous driving behavior in highway tunnels. Background Technology

[0002] With the increasing number of tunnels in operation, the travel distance on mountain roads has been greatly shortened and the traffic efficiency has been improved. However, due to the special characteristics of highway tunnels and the traffic environment, the safety management of tunnels has become the primary task of operation management.

[0003] Most traffic accidents in tunnels are caused by dangerous driving behavior. Currently, most existing technologies use image recognition methods to detect dangerous driving behavior. For example, patent application CN113239754A discloses a method and system for detecting and locating dangerous driving behavior in the Internet of Vehicles (IoV). This system analyzes dangerous driving behavior by extracting, segmenting, and matching features from real-time monitoring videos and images, extracts relevant dangerous driving behaviors, and matches them with preset alarm types. If a match is successful, an alarm is triggered via an audible and visual alarm module. However, this technical solution does not consider the driver's line of sight when issuing the alarm signal. It is possible that the alarm signal will only be issued after the driver has driven out of the line of sight, causing the driver to miss the alarm information. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes a proactive early warning method and system for highway tunnels based on dangerous driving behavior. This system can promptly issue alarm signals to prevent vehicles from going out of sight. The specific technical solution is as follows:

[0005] In a first aspect, a proactive early warning method for highway tunnels based on dangerous driving behavior is provided. In a first implementable mode of the first aspect, it includes:

[0006] The vehicle speed and driver parameters of the dangerously driven vehicle are obtained. The driver parameters include the driver's judgment time of the warning sign, visual half angle, and the height between the human eye and the highest point of the warning sign.

[0007] The minimum acceptable visibility distance for the driver is determined based on the driving speed and driver parameters. The visibility range of the warning sign is determined based on the minimum visibility distance and the position coordinates of the projection device. When a dangerously driven vehicle enters the visibility range, the projection device is controlled to project the warning sign.

[0008] In conjunction with the first implementable method of the first aspect, in the second implementable method of the first aspect, the driving speed and driver parameters of the dangerously driven vehicle are obtained, including:

[0009] Acquire the driving speed, following distance, lateral deviation, acceleration, and headlight characteristics of vehicles traveling in highway tunnels;

[0010] The vehicle speed, following distance, lateral offset, acceleration and / or headlight characteristics are compared with the corresponding judgment criteria, and the comparison results are used to determine whether the vehicle is engaging in dangerous driving behavior.

[0011] When a vehicle is found to be engaging in dangerous driving behavior, the vehicle will be designated as a dangerous driving vehicle.

[0012] In conjunction with the second feasible method of the first aspect, in the third feasible method of the first aspect, the driving speed, following distance, lateral offset, acceleration and headlight features of the driving vehicle are extracted from the video images of the highway tunnel.

[0013] In conjunction with the second feasible method of the first aspect, the fourth feasible method of the first aspect, in determining whether a vehicle is engaging in dangerous driving behavior, further includes:

[0014] Obtain the yaw rate and lateral and longitudinal acceleration of the moving vehicle;

[0015] Based on the vehicle's speed, following distance, lateral deviation, acceleration, yaw rate, and longitudinal and lateral acceleration, a trained support vector machine is used to determine whether a vehicle is engaging in dangerous driving behavior.

[0016] In conjunction with the first feasible method of the first aspect, in the fifth feasible method of the first aspect, obtaining the driver's interpretation time of the warning sign includes:

[0017] The total fixation time and saccade time of different drivers in the area of ​​interest were obtained through simulation experiments;

[0018] The corresponding normal distribution parameters are determined based on the sum of the fixation duration and saccade duration of all drivers;

[0019] The normal distribution quantile table is determined by using the normal distribution parameters, and the driver's interpretation time of the warning sign is determined by querying the normal distribution quantile table.

[0020] In conjunction with the first possible implementation of the first aspect, in the sixth possible implementation of the first aspect, an image positioning method is used to determine the location information of the dangerous driving vehicle, and the location information is used to determine whether the dangerous driving vehicle has entered the visible distance range.

[0021] In conjunction with the first implementable method of the first aspect, in the seventh implementable method of the first aspect, controlling the projection device to project warning signs includes:

[0022] Image recognition methods were used to determine the license plate information of vehicles driven dangerously.

[0023] The system generates corresponding warning signs by combining license plate information with the dangerous driving behavior of the vehicle.

[0024] Secondly, a proactive early warning system for highway tunnels based on dangerous driving behavior is provided. In a first feasible implementation of this second aspect, it includes:

[0025] The acquisition module is configured to acquire the driving speed of the dangerously driven vehicle and the driver parameters, including the driver's judgment time of the warning sign, visual half angle, and the height between the driver's eyes and the highest point of the warning sign;

[0026] The projection module is configured to determine the shortest visibility distance based on the vehicle speed and driver parameters, and to determine the visibility range of the warning sign based on the shortest visibility distance and the position coordinates of the projection device. When a dangerously driven vehicle enters the visibility range, the projection device is controlled to project the warning sign.

[0027] In conjunction with the first possible implementation of the second aspect, in the second possible implementation of the second aspect, the acquisition module includes:

[0028] The data acquisition unit is configured to acquire the sum of fixation time and saccade time of different drivers in the region of interest through simulation experiments;

[0029] The data statistics unit is configured to determine the corresponding normal distribution parameters based on the sum of the fixation time and saccade time of all drivers;

[0030] The parameter query unit is configured to determine a normal distribution quantile table through the normal distribution parameters, and to determine the driver's interpretation time of the warning sign by querying the normal distribution quantile table.

[0031] In conjunction with the first possible implementation of the second aspect, in the third possible implementation of the second aspect, the projection module uses an image positioning method to determine the location information of the dangerous driving vehicle, and determines whether the dangerous driving vehicle has entered the visible distance range based on the location information.

[0032] Beneficial effects: The active early warning method and system for highway tunnels based on dangerous driving behavior of the present invention can calculate the shortest visible distance at which the driver can receive the warning sign by combining the driving speed of the dangerous driving vehicle and the driver parameters. The visible distance range is determined by the shortest visible distance. The projection device can then project the warning sign after the dangerous driving vehicle enters the visible distance range and before it leaves the visible distance range, so as to remind the driver in time and avoid the driver missing the warning sign. Attached Figure Description

[0033] To more clearly illustrate the specific embodiments of the present invention, the accompanying drawings used in the specific embodiments will be briefly described below. In all the drawings, the elements or parts are not necessarily drawn to scale.

[0034] Figure 1 A flowchart of an active early warning method for highway tunnels based on dangerous driving behavior provided in an embodiment of the present invention;

[0035] Figure 2 This is a system block diagram of an active early warning system for highway tunnels based on dangerous driving behavior, provided in an embodiment of the present invention.

[0036] Figure 3 This is a diagram illustrating the driver's field of vision. Detailed Implementation

[0037] The embodiments of the technical solution of the present invention will now be described in detail with reference to the accompanying drawings. These embodiments are merely illustrative of the technical solution of the present invention and are therefore intended to limit the scope of protection of the present invention.

[0038] It should be understood that, in this embodiment, an image recognition system for detection is installed inside the highway tunnel. This image recognition system includes multiple high-definition cameras installed at intervals along the tunnel. To ensure no blind spots within the monitored area, high-definition cameras can be installed every 100-200 meters along the tunnel ceiling. When a vehicle enters the camera's monitoring range, the high-definition camera begins recording the vehicle's behavior.

[0039] It should also be understood that, in this embodiment, a projection system for projecting warning signs is also installed inside the highway tunnel. This projection system includes multiple projection devices, all of which can be installed sequentially at intervals. To avoid excessive visual load between the projection surface of the newly installed projection device and the existing traffic signs, leading to information omissions, a certain distance must be maintained between the projection position of the projection device and other traffic signs during installation. The specific calculation of this distance is as follows:

[0040] l = s + v * t² + jk;

[0041] Where j represents the judgment distance, k is the distance from the traffic sign after reading it, s represents the distance from the starting point to the traffic sign, v represents the vehicle speed, and t2 represents the time it takes for the vehicle operation to complete.

[0042] Furthermore, if the projection equipment is installed on a curve with a certain curvature, a multi-device fusion side wall projection method can be used. If the installation location is in a straight section inside the tunnel, a holographic projection method can be used to project the image above the tunnel.

[0043] like Figure 1The flowchart shown is for a proactive early warning method for highway tunnels based on dangerous driving behavior. This early warning method includes:

[0044] Step 1: Obtain the vehicle speed and driver parameters of the dangerously driven vehicle. These driver parameters include the driver's judgment time of the warning sign, visual half-angle, and the height between the driver's eye and the highest point of the warning sign.

[0045] Step 2: Determine the minimum acceptable visibility distance for the driver based on the driving speed and driver parameters, and determine the visibility range of the warning sign based on the minimum visibility distance and the projection position of the warning sign;

[0046] When a dangerously driven vehicle enters the visible range, the control projection equipment projects a warning sign.

[0047] Specifically, firstly, it can obtain the speed of the dangerously driven vehicle, as well as the driver's interpretation time of the warning sign, the visual half-angle of the driver's visual cone angle, and the height between the highest point of effective information of the warning sign and the human eye. For example... Figure 3 As shown, these driver parameters can all be obtained through preliminary simulation experiments. Then, based on the formation speed, interpretation time, visual half-angle, and height, the shortest visible distance at which the driver can receive the warning sign can be calculated. The specific calculation formula is as follows:

[0048]

[0049] Where S is the shortest visible distance, v is the vehicle speed, t1 is the interpretation time, h is the altitude, and α is the visual half-angle.

[0050] After calculating the shortest visible distance, the visible distance range of the warning sign can be determined based on the shortest visible distance and the projection position of the warning sign. In this embodiment, the visible distance range can be set to the lane range in the oncoming traffic direction where the distance between the warning sign's projection position and the lane is greater than the shortest visible distance. That is, when a dangerously driving vehicle travels to a position where the distance between it and the warning sign's projection position is greater than the shortest visible distance, the projection device can project the warning sign at that position. Conversely, if the dangerously driving vehicle passes a position where the distance between it and the warning sign's projection position is less than the shortest visible distance, the next projection device in the direction of the dangerously driving vehicle's travel can be controlled to project the warning sign at the corresponding position.

[0051] In this embodiment, optionally, obtaining the vehicle speed and driver parameters of the dangerously driven vehicle includes:

[0052] Acquire the driving speed, following distance, lateral deviation, acceleration, and headlight characteristics of vehicles traveling in highway tunnels;

[0053] The vehicle speed, following distance, lateral offset, acceleration and / or headlight characteristics are compared with the corresponding judgment criteria, and the comparison results are used to determine whether the vehicle is engaging in dangerous driving behavior.

[0054] When a vehicle is found to be engaging in dangerous driving behavior, the vehicle will be designated as a dangerous driving vehicle.

[0055] Specifically, when obtaining the speed and driver parameters of dangerously driven vehicles, it is necessary to first identify dangerously driven vehicles within highway tunnels. Specifically, firstly, the speed, following distance, lateral deviation, acceleration, and headlight characteristics of vehicles traveling in the highway tunnel can be obtained. Then, the speed, following distance, lateral deviation, acceleration, and headlight characteristics are compared with corresponding judgment criteria. The comparison results can then be used to determine whether the vehicle is a dangerously driven vehicle.

[0056] For example, comparing a vehicle's speed with a set speed limit threshold indicates dangerous driving behavior, and the vehicle can be identified as a dangerous driver. Similarly, comparing a vehicle's following distance with a safe distance threshold indicates that the vehicle is too close to the vehicle in front, and the vehicle can be identified as a dangerous driver. Driving trajectory can determine if a vehicle is crossing lane lines, while headlight characteristics, such as the intensity and position of the headlights, can determine if a vehicle is driving without its lights on.

[0057] Because dangerous driving behaviors also include fatigued driving and drunk driving, in order to more comprehensively and accurately determine whether a vehicle is engaged in dangerous driving, the yaw rate and lateral and longitudinal acceleration of the vehicle can be obtained. Based on the vehicle's speed, following distance, lateral deviation, acceleration, yaw rate, and lateral and longitudinal acceleration, a trained support vector machine can be used to determine whether a vehicle is engaging in dangerous driving behavior. Specifically, this includes:

[0058] First, based on real-vehicle driving experiments, vehicle data was acquired under abnormal driving conditions such as driver fatigue and distraction. The data collected by the front-end camera was preprocessed to obtain relevant index parameters. These parameters include: following distance *l*, lateral offset *c*, and lateral acceleration *a*. x Longitudinal acceleration a y These indicators are acquired every 3 seconds. The standard deviations σ(l), σ(c), and σ(a) of the four indicators for driving within a 60-second time window are calculated. x ), σ(a y ).

[0059] Then, to improve the accuracy of data calculation, min-max normalization is performed. The normalization expression is as follows:

[0060] y = (y max -y min )*(xx min ) / (x max -x min )+y min

[0061] The normalized data is then converted into a Support Vector Machine (SVM) data format, i.e., a CSV file. This file contains "f" to represent the category, which is either -1 or 1, where 1 indicates the vehicle is in a dangerous driving state and -1 indicates the vehicle is in a non-dangerous driving state; and "index_x:value" to represent the normalized vector values, expressed as a one-dimensional array. The data was divided into training and test sets in a 7:3 ratio.

[0062] Next, the hyperparameters and model structure of the support vector machine model are configured, and the Gaussian radial basis function is selected as the kernel function for support vector machine modeling. The specific expression of the model is as follows:

[0063]

[0064]

[0065] 0≤α i ≤C,i=1,2,...,m

[0066] Among them, kernel function

[0067] Regarding the selection of the penalty factor parameter C and the kernel function setting parameter σ, parameter search can be performed using methods such as grid search, genetic algorithm, or particle swarm optimization algorithm, and C and σ can be selected as the optimal parameters.

[0068] Then, the support vector machine is trained using the training set. The predicted values ​​are compared with the actual values, and the error function E is calculated. If the loss function E is less than the set value, the training is complete; otherwise, the relevant parameter values ​​are changed, and the optimal parameter values ​​of the penalty factor C and kernel function σ are determined again. At the same time, the optimal solution of α under these parameter values ​​is determined.

[0069] Then, choose one that satisfies optimal solution calculate

[0070]

[0071] Finally, based on the decision function expression regarding whether the driver is in a dangerous driving state, the system determines whether the vehicle is engaging in dangerous driving behavior. The specific decision function expression is as follows:

[0072]

[0073] In this embodiment, optionally, the vehicle speed, following distance, lateral offset, acceleration, and headlight features of the vehicle can be extracted from the video images of the highway tunnel. Specifically, the vehicle speed, following distance, lateral offset, acceleration, and headlight features of the vehicle can be extracted from the video images of the highway tunnel captured by a high-definition camera.

[0074] 1) Methods for extracting vehicle speed and acceleration:

[0075] Determine the time interval t between two adjacent photos in the video image of the highway tunnel captured by the camera, and the path distance s between these two photos. Then, calculate the vehicle speed V and acceleration a based on the path distance and the time interval. The specific calculation formula is as follows:

[0076]

[0077] 2) Following distance extraction method:

[0078] Extract images of the vehicles in motion, determine the installation spacing d of the road spikes on the tunnel side based on the images, and the number n of spikes between the two vehicles. Calculate the following distance based on the number of spikes and the installation distance, using the following formula:

[0079] l = nd.

[0080] 3) Method for extracting lateral offset:

[0081] Lane edge lines are identified and located by capturing inverted video images of highway tunnels using cameras. The lane centerline trajectory is then obtained by combining the coordinates of the lane edge lines. The vehicle's offset relative to the lane centerline is determined using the vehicle's coordinates on the image. If neither lane edge line is captured, the average offset displacement between consecutive images is taken as the vehicle's offset displacement for that image, to maintain the continuity of the trajectory.

[0082] 4) Methods for extracting vehicle headlight features:

[0083] A camera is used to photograph the headlights of moving vehicles. After image acquisition, noise reduction and image enhancement are performed. Then, feature extraction is conducted using methods such as edge detection and color analysis. Extracted features include brightness and beam direction. The features in the image are compared with standards to determine whether the vehicle headlights are turned on appropriately.

[0084] In this embodiment, optionally, in step 1, obtaining the driver's interpretation time of the warning sign includes:

[0085] The total fixation time and saccade time of different drivers in the area of ​​interest were obtained through simulation experiments;

[0086] The corresponding normal distribution parameters are determined based on the sum of the fixation duration and saccade duration of all drivers;

[0087] The normal distribution quantile table is determined by using the normal distribution parameters, and the driver's interpretation time of the warning sign is determined by querying the normal distribution quantile table.

[0088] Specifically, firstly, a tunnel driving simulation model of the same scale as the warning sign can be constructed. Multiple simulations are then conducted using this model to obtain the total fixation and saccade durations of multiple drivers within the region of interest. During the experiments, drivers can wear eye trackers while driving through the tunnel driving simulation model to obtain their total fixation and saccade durations within the region of interest. Due to the inherent error in eye tracker measurements, the region of interest can be calibrated to 1.1 times the size of the warning sign.

[0089] Then, statistical analysis can be performed on the sum of fixation and saccade durations for all drivers to determine the time it takes for drivers to interpret warning signs. Specifically, the normal distribution parameters can be determined by the sum of fixation and saccade durations for all drivers. The normal distribution function is as follows:

[0090]

[0091] Where μ is the driver's expected value of the sum of fixation duration and saccade duration within the area of ​​interest, and σ is the standard deviation of the sum of fixation duration and saccade duration.

[0092] The parameters of a random variable that follows a normal distribution function after standardization This allows us to determine the quantile table for the normal distribution. We then look up the value of u in the quantile table at a 75% confidence level, and the final interpretation time is (uσ+μ).

[0093] In this embodiment, optionally, an image positioning method is used to determine the location information of the dangerously driven vehicle, and the location information is used to determine whether the dangerously driven vehicle has entered the visible distance range.

[0094] Specifically, the location of a dangerously driven vehicle can be determined using video images of highway tunnels captured by high-definition cameras and existing image positioning technology. Based on the vehicle's location and the projected position of warning signs, the relative distance between the dangerously driven vehicle and the warning signs is determined. This relative distance is then compared to the shortest visible distance. If the relative distance is greater than the shortest visible distance, the dangerously driven vehicle is considered to have entered the visible distance range. Conversely, if the relative distance is less than the shortest visible distance, the dangerously driven vehicle is considered to have left the visible distance range.

[0095] In this embodiment, optionally, controlling the projection device to project warning signs includes:

[0096] Image recognition methods were used to determine the license plate information of vehicles driven dangerously.

[0097] The system generates corresponding warning signs by combining license plate information with the dangerous driving behavior of the vehicle.

[0098] Specifically, license plate information of dangerously driven vehicles can be identified using existing image recognition methods based on captured video images of highway tunnels. This license plate information, combined with the dangerous driving behavior, allows for the determination of appropriate warning signs, thereby directly alerting the relevant dangerous vehicles.

[0099] like Figure 2 The diagram shown is a system block diagram of a highway tunnel active warning system based on dangerous driving behavior. The warning system includes:

[0100] The acquisition module is configured to acquire the driving speed of the dangerously driven vehicle and the driver parameters, including the driver's judgment time of the warning sign, visual half angle, and the height between the driver's eyes and the highest point of the warning sign;

[0101] The projection module is configured to determine the shortest visibility distance based on the vehicle speed and driver parameters, and to determine the visibility range of the warning sign based on the shortest visibility distance and the projection position of the warning sign.

[0102] When a dangerously driven vehicle enters the visible range, the control projection equipment projects a warning sign.

[0103] Specifically, the warning system consists of an acquisition module and a projection module. The acquisition module can acquire the speed of the dangerously driven vehicle, the driver's interpretation time of the warning sign, the visual half-angle of the driver's visual cone angle, and the height between the driver's eye and the highest point of effective information of the warning sign. These driver parameters can be obtained through prior simulation experiments.

[0104] The projection module can calculate the minimum visible distance at which a driver can perceive a warning sign based on the formation speed, interpretation time, visual half-angle, and height. After calculating the minimum visible distance, the projection module can determine the visible range of the warning sign based on the minimum visible distance and the projection position of the warning sign. When a dangerously driven vehicle moves to a position where the distance between it and the projection position of the warning sign is greater than the minimum visible distance, the projection device can project the warning sign at that position.

[0105] In this embodiment, optionally, the acquisition module includes:

[0106] The data acquisition unit is configured to acquire the sum of fixation time and saccade time of different drivers in the region of interest through simulation experiments;

[0107] The data statistics unit is configured to determine the corresponding normal distribution parameters based on the sum of the fixation time and saccade time of all drivers;

[0108] The parameter query unit is configured to determine a normal distribution quantile table through the normal distribution parameters, and to determine the driver's interpretation time of the warning sign by querying the normal distribution quantile table.

[0109] Specifically, the acquisition module consists of a data acquisition unit, a data statistics unit, and a parameter query unit. The data acquisition unit acquires the total fixation duration and saccade duration of the driver within the region of interest, obtained through simulation experiments. The data statistics unit performs statistical analysis on the total fixation duration and saccade duration for all drivers to determine the normal distribution parameters of these totals. The parameter query unit obtains the driver's interpretation time by consulting a normal distribution quantile table.

[0110] In this embodiment, optionally, the projection module uses an image positioning method to determine the location information of the dangerous driving vehicle, and determines whether the dangerous driving vehicle has entered the visible distance range based on the location information.

[0111] Specifically, the projection module can locate the position of a dangerously driven vehicle using video images of a highway tunnel captured by a high-definition camera and existing image positioning technology. Based on the vehicle's position and the projected position of the warning sign, the relative distance between the dangerously driven vehicle and the warning sign is determined. This relative distance is compared to the shortest visible distance. If the relative distance is greater than the shortest visible distance, the dangerously driven vehicle has entered the visible distance range. Conversely, if the relative distance is less than the shortest visible distance, the dangerously driven vehicle has moved out of the visible distance range.

[0112] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for proactive early warning of dangerous driving behavior in highway tunnels, characterized in that, include: The vehicle speed and driver parameters of the dangerously driven vehicle are obtained. The driver parameters include the driver's judgment time of the warning sign, visual half angle, and the height between the human eye and the highest point of the warning sign. The minimum acceptable visibility distance for the driver is determined based on the driving speed and driver parameters, and the visibility range of the warning sign is determined based on the minimum visibility distance and the projection position of the warning sign. When a dangerously driven vehicle enters the visible range, control the projection equipment to project warning signs. The shortest visible distance at which a driver can perceive a warning sign is calculated based on vehicle speed, interpretation time, visual half-angle, and height. The specific calculation formula is as follows: ； For the shortest visible distance, For driving speed, To determine the reading time, For height, The visual half-angle is used to determine the interpretation time by summing the driver's fixation time and saccade time within the area of ​​interest.

2. The active early warning method for highway tunnels based on dangerous driving behavior according to claim 1, characterized in that, Obtain the vehicle speed and driver parameters of the dangerously driven vehicle, including: Acquire the driving speed, following distance, lateral deviation, acceleration, and headlight characteristics of vehicles traveling in highway tunnels; The vehicle speed, following distance, lateral offset, acceleration and / or headlight characteristics are compared with the corresponding judgment criteria, and the comparison results are used to determine whether the vehicle is engaging in dangerous driving behavior. When a vehicle is found to be engaging in dangerous driving behavior, the vehicle will be designated as a dangerous driving vehicle.

3. The active early warning method for highway tunnels based on dangerous driving behavior according to claim 2, characterized in that, The driving speed, following distance, lateral offset, acceleration, and headlight features of vehicles were extracted from video images of highway tunnels.

4. The active early warning method for highway tunnels based on dangerous driving behavior according to claim 2, characterized in that, Determining whether a vehicle is engaging in dangerous driving behavior also includes: Obtain the yaw rate and lateral and longitudinal acceleration of the moving vehicle; Based on the vehicle's speed, following distance, lateral deviation, acceleration, yaw rate, and longitudinal and lateral acceleration, a trained support vector machine is used to determine whether a vehicle is engaging in dangerous driving behavior.

5. The active early warning method for highway tunnels based on dangerous driving behavior according to claim 1, characterized in that, The time taken to obtain the driver's interpretation of the warning sign includes: The total fixation time and saccade time of different drivers in the area of ​​interest were obtained through simulation experiments; The corresponding normal distribution parameters are determined based on the sum of the fixation duration and saccade duration of all drivers; The normal distribution quantile table is determined by using the normal distribution parameters, and the driver's interpretation time of the warning sign is determined by querying the normal distribution quantile table.

6. The method for active early warning of highway tunnels based on dangerous driving behavior according to claim 1, characterized in that, The location information of the dangerously driven vehicle is determined by an image localization method, and the location information is used to determine whether the dangerously driven vehicle has entered the visible range.

7. The active early warning method for highway tunnels based on dangerous driving behavior according to claim 1, characterized in that, Warning signs for controlling projection equipment, including: Image recognition methods were used to determine the license plate information of vehicles driven dangerously. The system generates corresponding warning signs by combining license plate information with the dangerous driving behavior of the vehicle.

8. A highway tunnel active early warning system based on dangerous driving behavior, characterized in that, include: The acquisition module is configured to acquire the driving speed of the dangerously driven vehicle and the driver parameters, including the driver's judgment time of the warning sign, visual half angle, and the height between the driver's eyes and the highest point of the warning sign; The projection module is configured to determine the shortest visibility distance based on the vehicle speed and driver parameters, and to determine the visibility range of the warning sign based on the shortest visibility distance and the projection position of the warning sign. When a dangerously driven vehicle enters the visible range, control the projection equipment to project warning signs. The shortest visible distance at which a driver can perceive a warning sign is calculated based on vehicle speed, interpretation time, visual half-angle, and height. The specific calculation formula is as follows: ； For the shortest visible distance, For driving speed, To determine the reading time, For height, The visual half-angle is used to determine the interpretation time by summing the driver's fixation time and saccade time within the area of ​​interest.

9. The active early warning system for highway tunnels based on dangerous driving behavior according to claim 8, characterized in that, The acquisition module includes: The data acquisition unit is configured to acquire the sum of fixation time and saccade time of different drivers in the region of interest through simulation experiments; The data statistics unit is configured to determine the corresponding normal distribution parameters based on the sum of the fixation time and saccade time of all drivers; The parameter query unit is configured to determine a normal distribution quantile table through the normal distribution parameters, and to determine the driver's interpretation time of the warning sign by querying the normal distribution quantile table.

10. The active early warning system for highway tunnels based on dangerous driving behavior according to claim 8, characterized in that, The projection module uses an image positioning method to determine the location information of the dangerously driven vehicle, and determines whether the dangerously driven vehicle has entered the visible distance range based on the location information.

Images