Tunnel emergency stop lane induction emergency system and method

Abstract

The present application relates to the technical field of tunnel parking induction, and discloses a tunnel emergency parking strip induction emergency system and method, the system comprising a data acquisition module, an analysis and decision module, and an emergency indication module; wherein: the data acquisition module is used to collect environmental data and vehicle data in the tunnel; the analysis and decision module divides the tunnel into different emergency sections based on the environmental data, and sets an emergency indication strategy for each emergency section, and adjusts the emergency indication strategy based on vehicle data; the emergency indication module performs emergency induction on the emergency parking strip in the tunnel based on the emergency indication strategy; the analysis and decision module also performs abnormal vehicle detection based on vehicle data and generates an emergency induction strategy; the emergency indication module performs emergency parking induction on abnormal vehicles based on the emergency induction strategy; the present application can reasonably set and adjust the indication and induction modes of the emergency parking strip in the tunnel, which is conducive to ensuring tunnel traffic safety.

Description

Technical Field

[0001] The present application relates to the technical field of tunnel parking guidance, and in particular to a tunnel emergency parking lane guidance emergency system and method. Background Art

[0002] Currently, emergency stop lanes within tunnels have numerous shortcomings in terms of emergency guidance. Traditional methods for accident risk assessment rely primarily on simple accident statistics, lacking comprehensiveness and foresight. Some newly constructed tunnels, due to their short operating history, may lack sufficient accident data for a sound risk assessment, making it difficult to implement effective preventive measures. Such inaccurate risk assessments can lead to wasted resources or overlooked safety hazards.

[0003] Furthermore, existing emergency signaling strategies for tunnel emergency lanes are relatively limited. Most tunnels employ fixed lighting brightness and flashing frequencies, failing to dynamically adjust to actual traffic conditions and vehicle demand. Current technology lacks effective vehicle anomaly detection, hindering the timely identification of vehicles posing anomaly risks and the ability to proactively guide potential vehicles, resulting in unmanageable accident rates.

[0004] For example, patent application publication number CN115472028A discloses an intelligent warning and guidance method and system for emergency parking lanes in tunnels. The system includes a computing terminal, a parking lane, a display module, a voice reminder module, a rearview detection module, an event detection module, a person recognition module, a person analysis module, an emergency phone, a visual module, a tunnel broadcast, an indicator light, a flashing guidance sign, and a flashing light strip. This system can warn vehicles behind a vehicle entering a tunnel parking lane, display rear road conditions in real time, and effectively guide vehicles safely out of the tunnel parking lane.

[0005] For example, patent application publication number CN114613172A discloses an intelligent sensing warning system for a tunnel emergency stop lane, comprising a vehicle detection device, a yellow flashing guidance sign, a yellow flashing light strip, a full-color information board, an integrated telephone booth, an induction broadcast, a camera, and an in-tunnel information board. The vehicle detection device is located within the emergency stop lane, the yellow flashing guidance sign is located on the edge of the emergency stop lane pavement, the yellow flashing light strip is located on the side wall, the full-color information board is located at the front end, and the integrated telephone booth, induction broadcast, camera, and in-tunnel information board near the emergency stop lane are located at the rear end. This application can quickly detect vehicles within a tunnel emergency stop lane, effectively establish contact with faulty vehicles, and promptly issue warnings to vehicles behind vehicles entering or exiting the emergency stop lane.

[0006] The above technical solutions all have the problems raised by this background technology: lack of effective vehicle abnormality detection, inability to timely discover vehicles with abnormal risks, and inability to carry out induction planning for potential abnormal vehicles in advance, resulting in difficulty in controlling the accident rate.

[0007] The information disclosed in this background technology section is only intended to enhance the understanding of the overall background of the application and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art already known to ordinary technicians in this field. Summary of the Invention

[0008] The technical problem to be solved by this application is to overcome the defects of the existing technology and provide a tunnel emergency parking strip induction emergency system and method, so as to reasonably set and adjust the indication and induction mode of the emergency parking strip in the tunnel, which is conducive to ensuring tunnel traffic safety.

[0009] To solve the above technical problems, this application provides the following technical solutions:

[0010] On the one hand, the present application provides a tunnel emergency stop lane induction emergency system, including a data acquisition module, an analysis and decision module, and an emergency indication module; wherein:

[0011] The data acquisition module is used to collect environmental data and vehicle data in the tunnel;

[0012] The analysis and decision module divides the tunnel into different emergency sections based on the environmental data; the analysis and decision module is further configured to set an emergency indication strategy for each emergency section and adjust the emergency indication strategy based on the vehicle data;

[0013] The emergency indication module performs emergency guidance on the emergency parking strip in the tunnel based on the emergency indication strategy;

[0014] The analysis and decision module also performs abnormal vehicle detection based on vehicle data; when an abnormal vehicle is detected, the analysis and decision module generates an emergency induction strategy; and the emergency instruction module performs emergency stop induction on the abnormal vehicle based on the emergency induction strategy.

[0015] As a preferred solution of the tunnel emergency stop strip induction emergency system described in the present application, wherein: the environmental data includes distance data, curvature data, slope data, and accident frequency of each emergency stop strip in the tunnel;

[0016] The vehicle data includes the traffic volume in the tunnel, the speed of each vehicle, monitoring images, and temperature distribution diagrams;

[0017] The data acquisition module includes a data query unit, a traffic flow detection unit, a video monitoring unit, and a thermal imaging unit; the data query unit is used to obtain distance data, curvature data, slope data, and accident frequency of each emergency stop strip in the tunnel; wherein the distance data is obtained by obtaining position information of each emergency stop strip in the tunnel, and calculating the distance between each emergency stop strip and an adjacent emergency stop strip based on the position information as the distance data of each emergency stop strip;

[0018] The traffic flow detection unit is used to collect the traffic volume in the tunnel and the driving speed of each vehicle; the video monitoring unit is used to collect monitoring videos of each vehicle in the tunnel; and the thermal imaging unit is used to collect the temperature distribution map of each vehicle in the tunnel.

[0019] As a preferred solution of the tunnel emergency stop lane induction emergency system described in this application, the analysis and decision-making module includes an emergency strategy unit; the emergency strategy unit is used to divide the tunnel into different emergency sections and set an emergency indication strategy for each emergency section; specifically, it includes:

[0020] Encode the distance data, curvature data, slope data, and accident frequency of each emergency stop strip into a feature vector corresponding to the emergency stop strip;

[0021] Cluster the feature vectors of each emergency parking strip using a clustering algorithm; let the number of clustering results be N, and extract the emergency parking strips corresponding to the feature vectors in the N clustering results to form N emergency sections;

[0022] Calculate the average accident frequency of all emergency parking strips in each emergency strip; and classify the accident risk level for each emergency strip based on the average accident frequency.

[0023] An emergency indication strategy is set for each emergency section based on the accident risk level; the emergency indication strategy includes the light brightness and light flashing frequency of each emergency parking strip.

[0024] As a preferred solution of the tunnel emergency stop lane guidance emergency system described in this application, the analysis and decision module further includes a monitoring and adjustment unit; the monitoring and adjustment unit is used to adjust the emergency indication strategy, specifically including:

[0025] Determine the specific conditions of vehicle speed and traffic volume in the current tunnel;

[0026] If the current speed in the tunnel is high and the traffic volume is low, the brightness and flashing frequency of the lights in each emergency stop strip will be increased;

[0027] If the current speed and traffic volume in the tunnel are low, the brightness and flashing frequency of the lights in each emergency stop lane remain unchanged;

[0028] If the current speed in the tunnel is low and the traffic volume is high, the brightness and flashing frequency of the lights in each emergency stop strip will be reduced;

[0029] If the current vehicle speed and traffic volume in the tunnel are high, the light brightness and flashing frequency of each emergency stop strip will remain unchanged.

[0030] As a preferred solution of the tunnel emergency stop lane guidance emergency system described in this application, the method in which the monitoring and adjustment unit determines the specific conditions of the vehicle speed and traffic volume in the current tunnel is as follows:

[0031] The average speed of the vehicles in the tunnel is calculated based on the speed of each vehicle; the monitoring and adjustment unit is configured with a speed threshold, and if the average speed is greater than the speed threshold, it is determined that the current speed in the tunnel is high; otherwise, it is determined that the current speed in the tunnel is low;

[0032] The monitoring and adjusting unit is configured with a traffic flow threshold; if the traffic flow is higher than the traffic flow threshold, it is determined that the current traffic flow in the tunnel is high; otherwise, it is determined that the current traffic flow in the tunnel is low.

[0033] As a preferred solution of the tunnel emergency stop strip induction emergency system described in the present application, wherein: the emergency indication module includes a lighting indication unit and a dimming drive unit; wherein, the lighting indication unit includes an LED lamp installed in each emergency stop strip; the LED lamp is used to indicate the position of the corresponding emergency stop strip;

[0034] The dimming drive unit includes a control device matched with each LED lamp; the control device is used to execute the emergency indication strategy and control the light brightness and light flashing frequency of each LED lamp in the emergency parking zone.

[0035] As a preferred solution of the tunnel emergency stop lane guidance emergency system described in this application, the analysis and decision-making module further includes an anomaly detection unit, which is used to detect abnormal vehicles; the anomaly detection unit is configured with a target detection algorithm and a multi-target tracking algorithm; the method for detecting abnormal vehicles is as follows:

[0036] Continuously acquiring monitoring images of each vehicle in the tunnel from the video monitoring unit;

[0037] Detecting vehicles from the monitoring image using a target detection algorithm; tracking and detecting each vehicle in the continuous monitoring image using a multi-target tracking algorithm;

[0038] Extract two frames of surveillance images to perform random inspections on any vehicle and calculate the vehicle's inspection speed; continue to inspect each vehicle, and if the difference between the inspection speed and the average driving speed of any vehicle in m consecutive inspections exceeds a preset error range, the vehicle is marked as a risky vehicle;

[0039] Obtain a thermal image of the risk vehicle based on a thermal imaging unit; obtain the position coordinates of the risk vehicle; and obtain curvature data and slope data corresponding to the position coordinates of the risk vehicle;

[0040] The anomaly detection unit is also configured with an anomaly detection model, which inputs the average driving speed and the sampling speed of any risk vehicle, the thermal image, the curvature data and the slope data corresponding to the position coordinates of the risk vehicle into the anomaly detection model and outputs the abnormal vehicle detection result.

[0041] As a preferred solution of the tunnel emergency stop lane induction emergency system described in this application, the analysis and decision-making module further includes an emergency induction unit; the emergency induction unit is used to generate an emergency induction strategy; specifically, it includes:

[0042] Acquire the position coordinates of the abnormal vehicle; select the emergency parking strip closest to the abnormal vehicle in the direction of the abnormal vehicle's advance as the induced parking strip; the emergency induced unit sends an emergency induced instruction to the dimming drive unit, and the dimming drive unit responds to the emergency induced instruction to increase the light brightness and light flashing frequency of the induced parking strip.

[0043] As a preferred solution of the tunnel emergency parking strip induction emergency system described in the present application, wherein: the emergency indication module also includes an induction identification unit; the emergency induction strategy also includes: the emergency induction unit sends a parking induction instruction to the induction identification unit; the induction identification unit responds to the parking induction instruction and performs emergency parking induction on the abnormal vehicle; the induction identification unit includes a display screen installed along the tunnel; the emergency parking induction for the abnormal vehicle includes: indicating the direction of the induction parking strip by an arrow on the display screen, and marking the distance of the induction parking strip from the display screen.

[0044] In a second aspect, the present application provides a method for inducing emergency response in an emergency stop zone in a tunnel, specifically comprising: collecting environmental data in the tunnel; dividing the tunnel into different emergency sections based on the environmental data, and setting an emergency indication strategy for each emergency section;

[0045] Collecting vehicle data in the tunnel; adjusting the emergency indication strategy based on the vehicle data;

[0046] Provide emergency guidance for emergency parking strips in tunnels based on emergency indication strategies;

[0047] Identify risky vehicles in tunnels based on vehicle data; perform anomaly detection on risky vehicles and identify abnormal vehicles;

[0048] Acquire the position coordinates of the abnormal vehicle; generate an emergency induction strategy for the abnormal vehicle based on the position coordinates, and perform emergency stop induction on the abnormal vehicle according to the emergency induction strategy.

[0049] Compared with the prior art, the beneficial effects achieved by this application are as follows:

[0050] By analyzing environmental data such as the distance, curvature, slope, and accident frequency of each emergency stop lane in the tunnel, the tunnel is divided into different emergency sections and the accident risk levels are reasonably divided. This avoids the one-sidedness of relying solely on accident risk data, can better predict potential high-accident areas, and provide a basis for the subsequent formulation of emergency instruction strategies.

[0051] The emergency sign strategy for each emergency section is dynamically adjusted based on the number of vehicles. When speeds are high and traffic volume is low, the brightness and frequency are increased to facilitate quick vehicle attention. When speeds are low and traffic volume is high, the brightness and frequency are reduced to reduce the visual burden on drivers, making the emergency signs more aligned with actual traffic conditions and improving guidance effectiveness.

[0052] By utilizing target detection algorithms, multi-target tracking algorithms, and anomaly detection models, combined with vehicle monitoring images, spot check speeds, thermal imaging images, and location-related curvature and slope data, the system can accurately detect abnormal vehicles. By increasing the brightness and flashing frequency of the lights in the induced parking lanes and providing induced information on the display screen, it can effectively guide abnormal vehicles to the nearest emergency parking lane, thereby ensuring tunnel traffic safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] To more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present application. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0054] Figure 1 A schematic diagram of the structure of a tunnel emergency stop strip induction emergency system provided in this application;

[0055] Figure 2 A functional schematic diagram of a tunnel emergency stop strip induction emergency system provided in this application;

[0056] Figure 3 This is a flow chart of the method for dividing emergency sections and setting emergency indication strategies provided in this application. DETAILED DESCRIPTION

[0057] The technical solution of the present application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations on the technical solution of the present application. Unless there is a conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0058] Example 1

[0059] This embodiment introduces a tunnel emergency stop lane guidance emergency system. Figure 1 The system includes data acquisition module, analysis and decision module, and emergency instruction module; the functions of each module are as follows: Figure 2 shown.

[0060] The data acquisition module is used to collect environmental data and vehicle data in the tunnel;

[0061] The environmental data includes distance data, curvature data, slope data, and accident frequency of each emergency stop strip in the tunnel;

[0062] The vehicle data includes the traffic volume in the tunnel, the speed of each vehicle, monitoring images, and temperature distribution diagrams;

[0063] The data acquisition module includes a data query unit, a traffic flow detection unit, a video surveillance unit, and a thermal imaging unit. The data query unit is used to obtain distance data, curvature data, slope data, and accident frequency for each emergency stop lane in the tunnel. The distance data is obtained by obtaining the location information of each emergency stop lane in the tunnel and, based on the location information, calculating the distance between each emergency stop lane and its adjacent emergency stop lane as the distance data for each emergency stop lane. For the emergency stop lane closest to either end of the tunnel, it has only one adjacent emergency stop lane, and the distance between the two adjacent emergency stop lanes serves as its distance data. For any emergency stop lane located in the middle of the tunnel, it has adjacent emergency stop lanes on both sides. The distances between the two adjacent emergency stop lanes are calculated and averaged to serve as the distance data. Distance data is one indicator of the importance of each emergency stop lane. Emergency stop lanes with larger distance data experience greater parking pressure when a vehicle needs to make an emergency stop nearby due to the greater distance of other nearby emergency stop lanes.

[0064] The curvature data is the radius of curvature at the location of the emergency stop strip; the slope data is the slope of the location of the emergency stop strip, with positive values ​​representing uphill and negative values ​​representing downhill. Historical accident records within the tunnel are obtained, including the location of the accident (specifically, a certain location or section within the tunnel). Each emergency stop strip is divided into associated intervals, such as the range within 10 meters on both sides of the emergency stop strip as its associated interval. Statistical analysis of historical accident records is performed to calculate the accident frequency within the associated interval of each emergency stop strip. For example, the number of accidents within the associated interval within a year or a quarter is the accident frequency of the corresponding emergency stop strip.

[0065] The traffic flow detection unit is used to collect the traffic volume in the tunnel and the driving speed of each vehicle; the traffic flow detection unit is equipped with geomagnetic sensors, microwave radars and other equipment installed above or on both sides of the lanes in the tunnel, which can accurately detect the traffic volume in the tunnel and the driving speed of each vehicle.

[0066] The video surveillance unit is used to collect surveillance video of each vehicle in the tunnel; the video surveillance unit includes a high-definition camera installed on the top or side wall of the tunnel. The monitoring range covers the entire tunnel and all emergency parking lanes, and the real-time location of each vehicle can be obtained through the monitoring screen.

[0067] The thermal imaging unit is used to collect temperature distribution maps for each vehicle in the tunnel. The thermal imaging unit includes a thermal imaging camera installed near each emergency stop. It detects the temperature distribution of each vehicle within the associated area of ​​the emergency stop and generates a temperature distribution map for each vehicle.

[0068] The analysis and decision module divides the tunnel into different emergency sections based on the environmental data; the analysis and decision module is further configured to set an emergency indication strategy for each emergency section and adjust the emergency indication strategy based on the vehicle data;

[0069] Each emergency stop zone has different environmental data, resulting in different emergency stopping requirements. Therefore, emergency indication strategies are tailored to each emergency stop zone based on specific environmental data. For example, emergency stops are often required at curves or accident-prone areas, so enhanced lighting and guidance are needed for emergency stop zones in these locations.

[0070] The analysis and decision-making module includes an emergency strategy unit and a monitoring and adjustment unit; wherein:

[0071] The emergency strategy unit is used to divide the tunnel into different emergency sections and set an emergency indication strategy for each emergency section. Figure 3 , as follows:

[0072] Performing data cleaning on the environmental data, and normalizing the distance data, curvature data, slope data, and accident frequency respectively;

[0073] Encode the distance data, curvature data, slope data, and accident frequency of each emergency stop strip into a feature vector corresponding to the emergency stop strip;

[0074] A clustering algorithm is used to cluster the feature vectors of each emergency stop strip. Let N be a positive integer. The emergency stop strips corresponding to the feature vectors in each of the N clusters are extracted to form N emergency sections. The feature vectors of the emergency stop strips in any emergency section all correspond to the same cluster. Based on the data characteristics and analysis objectives, the K-Means clustering algorithm can be used. The elbow rule can be used to determine the appropriate number of clusters, i.e., the number of clusters, and a curve is plotted showing the relationship between the number of clusters and the sum of squared errors (SSE) within the clusters. As the number of clusters increases, the SSE decreases. When the curve begins to decrease at a certain number of clusters, this value can be used as the initial setting. Emergency stop strips are divided into different clusters, each with similar topography and accident characteristics. For example, the emergency stop strips in one cluster may all be located in long, straight tunnel sections with a low accident frequency, while the emergency stop strips in another cluster may all be located on curves with a high accident frequency.

[0075] Calculate the mean accident frequency of all emergency stop strips in each emergency strip; assign an accident risk level to each emergency strip based on the mean accident frequency; for example, if the clustering result data N is 3, three emergency segments are generated; among the three emergency segments, the segment with the highest mean accident frequency has a high risk level, the segment with the second highest mean accident frequency has a medium risk level, and the segment with the lowest mean accident frequency has a low risk level;

[0076] Relying solely on accident risk data may not fully reflect the characteristics of different areas within the tunnel. Some newer tunnels may not have enough data to make a reasonable risk level classification, or there may be deviations in the accident data in certain areas due to accidental factors. For example, several abnormal accidents occurred near an emergency parking strip in a short period of time (such as sudden vehicle failures or sudden illness of the driver), but from a long-term and actual road condition perspective, this area is not a high-risk area. Therefore, this application selected distance data, curvature data, and slope data that are highly correlated with the frequency of accidents, and clustered them together with the frequency of accidents to better predict potential high-accident areas.

[0077] An emergency indication strategy is set for each emergency section based on the accident risk level. This strategy includes the lighting brightness and flashing frequency for each emergency stop strip. For example, for an emergency section with a low accident risk level, the lighting brightness for each emergency stop strip indicating the parking position is 40 to 60 lumens, with a flashing frequency of 10 to 20 times per minute. The brightness and flashing frequency can be relatively low, providing basic indication illumination without affecting the driver's normal vision. For an emergency section with a medium accident risk level, the lighting brightness for each emergency stop strip is 60 to 80 lumens, with a flashing frequency of 30 to 50 times per minute. This medium brightness and flashing frequency allows drivers to effectively identify the facilities and signs within the emergency stop strip. For an emergency section with a low accident risk level, the lighting brightness for each emergency stop strip is 80 to 100 lumens, with a flashing frequency of 60 to 90 times per minute. Setting a higher brightness ensures that the emergency stop strip is clearly visible under various lighting conditions, helping drivers better judge the distance and location of the emergency stop strip. The high-frequency flashing light mode allows drivers to notice the location of the emergency parking lane from a long distance, making it easier for them to prepare for parking in advance.

[0078] The monitoring adjustment unit is used to adjust the emergency indication strategy, specifically including:

[0079] Determine the specific conditions of vehicle speed and traffic volume in the current tunnel;

[0080] The average speed of the vehicles in the tunnel is calculated based on the speed of each vehicle; the monitoring and adjustment unit is configured with a speed threshold, and if the average speed is greater than the speed threshold, it is determined that the current speed in the tunnel is high; otherwise, it is determined that the current speed in the tunnel is low;

[0081] The monitoring and adjustment unit is configured with a traffic flow threshold; if the traffic flow is higher than the traffic flow threshold, it is determined that the traffic flow in the current tunnel is high; otherwise, it is determined that the traffic flow in the current tunnel is low;

[0082] If the current vehicle speed in the tunnel is high and the traffic volume is low, the light brightness and flashing frequency of each emergency stop strip will be increased; for example, the light brightness and flashing frequency will be increased by 20% to facilitate fast-moving vehicles to pay attention to the emergency stop strip in time.

[0083] If the current vehicle speed and traffic volume in the tunnel are low, the brightness and flashing frequency of the lights in each emergency stop lane will remain unchanged. At this time, the traffic conditions in the tunnel are good, and the driver has enough time and vision to observe the location of the emergency stop lane, so there is no need to adjust the emergency indication strategy.

[0084] If the current vehicle speed in the tunnel is low and the traffic volume is high, the light brightness and flashing frequency of each emergency stop strip will be reduced; for example, the light brightness and flashing frequency will be reduced by 10%; higher light brightness will increase the visual burden on the driver; reducing the light flashing frequency can also reduce the stimulation to drivers in congested and slow-moving situations.

[0085] If the current vehicle speed and traffic volume in the tunnel are high, the brightness and flashing frequency of the lights in each emergency stop lane will remain unchanged; when the vehicle speed is high, the emergency stop lane should be made more eye-catching, but in the case of high traffic volume, in order to prevent excessive brightness from causing glare that interferes with the driver's vision and to prevent high-frequency flashing from causing visual confusion, the emergency indication strategy will not be adjusted.

[0086] The emergency indication module performs emergency guidance on the emergency parking strip in the tunnel based on the emergency indication strategy;

[0087] The emergency indication module includes a lighting indication unit and a dimming drive unit; wherein, the lighting indication unit includes an LED lamp installed in each emergency parking strip; the LED lamp is used to indicate the position of the corresponding emergency parking strip; each LED lamp has an independent control function and can flexibly adjust the brightness and flashing frequency.

[0088] The dimming drive unit includes a control device associated with each LED lamp. This control device is used to implement the emergency indication strategy and control the brightness and flashing frequency of each LED lamp in the emergency stop zone. The control device can precisely control the current and voltage of the corresponding LED lamp, achieving rapid and stable adjustment of the lamp's brightness and flashing pattern.

[0089] The analysis and decision module also performs abnormal vehicle detection based on vehicle data; when an abnormal vehicle is detected, the analysis and decision module generates an emergency induction strategy; and the emergency instruction module performs emergency stop induction on the abnormal vehicle based on the emergency induction strategy.

[0090] The analysis and decision module also includes an anomaly detection unit and an emergency induction unit; wherein the anomaly detection unit is used to detect abnormal vehicles; the anomaly detection unit is configured with a target detection algorithm and a multi-target tracking algorithm; the method for detecting abnormal vehicles is as follows:

[0091] Continuously acquiring monitoring images of each vehicle in the tunnel from the video monitoring unit;

[0092] Vehicles are detected from the surveillance images using an object detection algorithm, and each detected vehicle is numbered. Deep learning-based object detection algorithms, such as the YOLO series of models and Faster R-CNN, are pre-trained on large-scale datasets and fine-tuned on task-specific datasets to accurately identify vehicles in images.

[0093] Through a multi-target tracking algorithm, each vehicle is tracked and detected in continuous monitoring images based on the number; multi-target tracking algorithms such as DeepSORT can achieve continuous tracking of vehicles in monitoring images of different frames based on appearance features (such as the color and shape of the vehicle).

[0094] Extract two surveillance images and spot-check any vehicle, calculating the vehicle's spot-check speed. Continue spot-checking each vehicle. If the difference between the spot-check speed and the average speed for m consecutive spot checks exceeds a preset error range, the vehicle is marked as a risky vehicle; m is a positive integer. The object detection algorithm provides bounding box information for each vehicle, which can be used to determine the vehicle's two-dimensional coordinates in the surveillance image. Using the surveillance camera's intrinsic parameters (such as focal length, principal point coordinates, etc.) and extrinsic parameters (such as rotation matrices and translation vectors), coordinate transformation can be performed to convert the two-dimensional coordinates in the surveillance image to the vehicle's actual position coordinates within the tunnel. The difference (i.e., distance) between the position coordinates of the same vehicle in the two surveillance images is calculated and divided by the time difference between the two surveillance images to obtain the spot-check speed of the corresponding vehicle.

[0095] Obtain a thermal image of the risk vehicle based on a thermal imaging unit; obtain the position coordinates of the risk vehicle, and obtain curvature data and slope data corresponding to the position coordinates of the risk vehicle;

[0096] The anomaly detection unit is also configured with an anomaly detection model, which inputs the average driving speed and the sampling speed of any risk vehicle, the thermal image, the curvature data and the slope data corresponding to the position coordinates of the risk vehicle into the anomaly detection model and outputs the abnormal vehicle detection result.

[0097] The anomaly detection model includes an input layer, a feature extraction layer, a feature combination layer, and an output layer;

[0098] The input layer is used to receive the average driving speed and the sampling speed of any risk vehicle, the thermal image, the curvature data and the slope data corresponding to the position coordinates of the risk vehicle as model input; the input layer includes an image input layer and a numerical input layer, wherein the image input layer is used to receive the thermal image, which is cropped and resized, with a length and width of 224 pixels; the numerical input layer is used to receive the average driving speed, sampling speed, curvature and slope, and these values ​​are encoded and input into the numerical input layer in the form of feature vectors.

[0099] The feature extraction layer extracts a first feature based on the thermal image of the risky vehicle and a second feature based on the average driving speed, the inspection speed of the risky vehicle, and the curvature and slope data corresponding to the risky vehicle's location coordinates. The feature extraction layer includes an image feature extraction component and a numerical feature extraction component. The image feature extraction component is composed of multiple convolutional layers and pooling layers, which gradually extract the features of the thermal image as the first feature through convolution and pooling operations and flatten the first feature. The numerical feature extraction component includes multiple fully connected layers, which map the input feature vector to a high-dimensional space as the second feature.

[0100] The feature combination layer concatenates the first feature and the second feature to obtain a fused feature; the concatenation of the first feature and the second feature enables the model to comprehensively consider multiple types of data features.

[0101] The output layer maps the fused features to an abnormal probability value for a risky vehicle. Using Sigmoid as the activation function, the output layer maps the concatenated fused features to an output value between 0 and 1, representing the abnormal probability of the risky vehicle. An abnormality threshold is set based on actual needs. If the abnormal probability threshold for any risky vehicle exceeds the threshold, the risky vehicle is marked as an abnormal vehicle.

[0102] Each vehicle's thermal image is captured by a thermal imaging camera located in front of the vehicle in its forward direction, ensuring that thermal images from different vehicles are collected in the same direction and are comparable. Thermal images provide information about the vehicle's temperature distribution. Normal vehicles have a normal temperature distribution. For example, components such as the engine, brakes, and tires operate within a certain temperature range. If a thermal image shows an abnormally high or low temperature for a component, this may indicate a fault. Leveraging the characteristics of convolutional neural networks, features are automatically extracted from thermal images that may indicate a vehicle fault. By processing numerical data through multiple fully connected layers, the relationships between these numerical features can be learned. For example, data features such as vehicle speed at different curvatures and slopes, and the consistency between the average speed of risky vehicles and that of tunnels, can be mined to detect potential anomalies.

[0103] The emergency induction unit is used to generate an emergency induction strategy; specifically, it includes:

[0104] Obtain the position coordinates of the abnormal vehicle; select the emergency parking strip closest to the abnormal vehicle in the direction of the abnormal vehicle's advance as the induced parking strip; the emergency induction unit sends an emergency induction instruction to the dimming drive unit, and the dimming drive unit responds to the emergency induction instruction to increase the light brightness and light flashing frequency of the induced parking strip; for example, set a maximum light brightness and a maximum light flashing frequency for each emergency parking strip, and when the dimming drive unit receives the emergency induction instruction, switch the light brightness of the corresponding emergency parking strip to the maximum light brightness, and switch the light flashing frequency to the maximum light flashing frequency.

[0105] The emergency indication module also includes an induction identification unit; the emergency induction strategy also includes: the emergency induction unit sends a parking induction instruction to the induction identification unit; the induction identification unit responds to the parking induction instruction and performs emergency parking induction on the abnormal vehicle; the induction identification unit includes a display screen installed along the tunnel; the emergency parking induction for the abnormal vehicle includes: indicating the direction of the induction parking strip by an arrow on the display screen, and marking the distance of the induction parking strip from the display screen; by arranging high-brightness, high-contrast display screens near each emergency parking strip and at key positions along the tunnel, it is ensured that abnormal vehicles can be clearly and accurately guided to the emergency parking strip closest to them in various environments.

[0106] Example 2

[0107] This embodiment is the second embodiment of the present application. Based on the same inventive concept as the first embodiment, this embodiment introduces a tunnel emergency stop lane induction emergency method, specifically including:

[0108] Collecting environmental data within the tunnel; specifically, collecting distance data, curvature data, slope data, and accident frequency of each emergency stop strip in the tunnel; dividing the tunnel into different emergency sections based on the environmental data, and setting an emergency indication strategy for each emergency section; the emergency indication strategy includes the light brightness and light flashing frequency of each emergency stop strip.

[0109] Collecting vehicle data in the tunnel; adjusting the emergency indication strategy based on the vehicle data; and adjusting the emergency indication strategy based on specific conditions of vehicle speed and traffic volume in the tunnel.

[0110] Emergency guidance is carried out for the emergency parking strips in the tunnel based on the emergency indication strategy; the working parameters of the LED lamps in each emergency parking strip are adjusted according to the light brightness and light flashing frequency specified by the emergency indication strategy, thereby realizing personalized emergency guidance for each emergency parking strip.

[0111] Identify risky vehicles in the tunnel based on vehicle data; perform anomaly detection on these vehicles and identify abnormal vehicles. Use target detection and multi-target tracking algorithms to determine whether the speed of any detected vehicle is synchronized with the average speed of all vehicles. Determine if the speed is out of sync with the average speed of all vehicles. Use a trained anomaly detection model to determine whether each risky vehicle is an abnormal vehicle.

[0112] The system obtains the location coordinates of the abnormal vehicle; generates an emergency guidance strategy for the abnormal vehicle based on the location coordinates, and guides the abnormal vehicle to an emergency stop according to the emergency guidance strategy. The display screen indicates the direction and distance of the nearest emergency stop lane to the abnormal vehicle, thereby accurately guiding the abnormal vehicle to the nearest emergency stop lane.

[0113] The specific functions of the above steps are realized by referring to the relevant contents of the tunnel emergency stop lane induction emergency system described in Example 1, which will not be described in detail.

[0114] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0115] The above describes the embodiments of the present application in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also make many forms without departing from the purpose and scope of protection of this application, all of which are protected by this application.

Claims

1. A tunnel emergency stop lane induction emergency system, characterized by: It includes data acquisition module, analysis and decision module, and emergency instruction module; among which: The data acquisition module is used to collect environmental data and vehicle data in the tunnel; The analysis and decision module divides the tunnel into different emergency sections based on the environmental data; the analysis and decision module is further configured to set an emergency indication strategy for each emergency section and adjust the emergency indication strategy based on the vehicle data; The analysis and decision module includes an emergency strategy unit; the emergency strategy unit is used to divide the tunnel into different emergency sections and set an emergency indication strategy for each emergency section; specifically, it includes: Encode the distance data, curvature data, slope data, and accident frequency of each emergency stop strip into a feature vector corresponding to the emergency stop strip; Cluster the feature vectors of each emergency parking strip using a clustering algorithm; let the number of clustering results be N, and extract the emergency parking strips corresponding to the feature vectors in the N clustering results to form N emergency sections; Calculate the average accident frequency of all emergency parking strips in each emergency strip; and classify the accident risk level for each emergency strip based on the average accident frequency. Setting an emergency indication strategy for each emergency section based on the accident risk level; the emergency indication strategy includes the light brightness and light flashing frequency of each emergency parking strip; The emergency indication module performs emergency guidance on the emergency parking strip in the tunnel based on the emergency indication strategy; The analysis and decision module also performs abnormal vehicle detection based on vehicle data; when an abnormal vehicle is detected, the analysis and decision module generates an emergency induction strategy; and the emergency instruction module performs emergency stop induction on the abnormal vehicle based on the emergency induction strategy.

2. The tunnel emergency stop lane guidance emergency system according to claim 1, characterized in that: The environmental data includes distance data, curvature data, slope data, and accident frequency of each emergency stop strip in the tunnel; The vehicle data includes the traffic volume in the tunnel, the speed of each vehicle, monitoring images, and temperature distribution diagrams; The data acquisition module includes a data query unit, a traffic flow detection unit, a video monitoring unit, and a thermal imaging unit; the data query unit is used to obtain distance data, curvature data, slope data, and accident frequency of each emergency stop strip in the tunnel; wherein the distance data is obtained by obtaining position information of each emergency stop strip in the tunnel, and calculating the distance between each emergency stop strip and an adjacent emergency stop strip based on the position information as the distance data of each emergency stop strip; The traffic flow detection unit is used to collect the traffic volume in the tunnel and the driving speed of each vehicle; the video monitoring unit is used to collect monitoring videos of each vehicle in the tunnel; and the thermal imaging unit is used to collect the temperature distribution map of each vehicle in the tunnel.

3. The tunnel emergency stop lane guidance emergency system according to claim 2, characterized in that: The analysis and decision module further includes a monitoring and adjustment unit; the monitoring and adjustment unit is used to adjust the emergency instruction strategy, specifically including: Determine the specific conditions of vehicle speed and traffic volume in the current tunnel; If the current speed in the tunnel is high and the traffic volume is low, the brightness and flashing frequency of the lights in each emergency stop strip will be increased; If the current speed and traffic volume in the tunnel are low, the brightness and flashing frequency of the lights in each emergency stop lane remain unchanged; If the current speed in the tunnel is low and the traffic volume is high, the brightness and flashing frequency of the lights in each emergency stop strip will be reduced; If the current vehicle speed and traffic volume in the tunnel are high, the light brightness and flashing frequency of each emergency stop strip will remain unchanged.

4. The tunnel emergency stop lane guidance emergency system according to claim 3, characterized in that: The method for the monitoring and adjustment unit to determine the specific conditions of the vehicle speed and traffic volume in the current tunnel is as follows: The average speed of the vehicles in the tunnel is calculated based on the speed of each vehicle; the monitoring and adjustment unit is configured with a speed threshold, and if the average speed is greater than the speed threshold, it is determined that the current speed in the tunnel is high; otherwise, it is determined that the current speed in the tunnel is low; The monitoring and adjustment unit is configured with a vehicle flow threshold; If the traffic volume is higher than the traffic volume threshold, it is determined that the traffic volume in the current tunnel is high; otherwise, it is determined that the traffic volume in the current tunnel is low.

5. The tunnel emergency stop lane guidance emergency system according to claim 4, characterized in that: The emergency indication module includes a lighting indication unit and a dimming drive unit; wherein the lighting indication unit includes an LED lamp installed in each emergency parking strip; the LED lamp is used to indicate the position of the corresponding emergency parking strip; The dimming drive unit includes a control device matched with each LED lamp; the control device is used to execute the emergency indication strategy and control the light brightness and light flashing frequency of each LED lamp in the emergency parking zone.

6. The tunnel emergency stop lane guidance emergency system according to claim 5, characterized in that: The analysis and decision module further includes an anomaly detection unit, which is used to detect abnormal vehicles. The anomaly detection unit is configured with a target detection algorithm and a multi-target tracking algorithm. The method for detecting abnormal vehicles is as follows: Continuously acquiring monitoring images of each vehicle in the tunnel from the video monitoring unit; Detecting vehicles from the monitoring image using a target detection algorithm; tracking and detecting each vehicle in the continuous monitoring image using a multi-target tracking algorithm; Extract two frames of surveillance images to perform random inspections on any vehicle and calculate the vehicle inspection speed; Continuously perform random inspections on each vehicle. If the difference between the inspection speed and the average driving speed of any vehicle exceeds a preset error range during m consecutive random inspections, the vehicle will be marked as a risky vehicle. Obtain a thermal image of the risk vehicle based on a thermal imaging unit; obtain the position coordinates of the risk vehicle; and obtain curvature data and slope data corresponding to the position coordinates of the risk vehicle; The anomaly detection unit is also configured with an anomaly detection model, which inputs the average driving speed and the sampling speed of any risk vehicle, the thermal image, the curvature data and the slope data corresponding to the position coordinates of the risk vehicle into the anomaly detection model and outputs the abnormal vehicle detection result.

7. The tunnel emergency stop lane guidance emergency system according to claim 6, characterized in that: The analysis and decision module further includes an emergency induction unit; the emergency induction unit is used to generate an emergency induction strategy; specifically, it includes: Obtain the position coordinates of the abnormal vehicle; select the emergency parking strip closest to the abnormal vehicle in the direction of the abnormal vehicle's advance as the induced parking strip; the emergency induction unit sends an emergency induction instruction to the dimming drive unit, and the dimming drive unit responds to the emergency induction instruction to increase the light brightness and light flashing frequency of the induced parking strip.

8. The tunnel emergency stop lane guidance emergency system according to claim 7, characterized in that: The emergency indication module also includes an induction identification unit; the emergency induction strategy also includes: the emergency induction unit sends a parking induction instruction to the induction identification unit; the induction identification unit responds to the parking induction instruction and performs emergency parking induction on the abnormal vehicle; the induction identification unit includes a display screen installed along the tunnel; the emergency parking induction for the abnormal vehicle includes: indicating the direction of the induction parking strip by an arrow on the display screen, and marking the distance of the induction parking strip from the display screen.

9. A tunnel emergency stop lane induction emergency method, which is implemented based on a tunnel emergency stop lane induction emergency system according to any one of claims 1 to 8, characterized in that: The following steps are involved: Collect environmental data inside the tunnel; Dividing the tunnel into different emergency sections based on the environmental data, and setting an emergency indication strategy for each emergency section; Collecting vehicle data in the tunnel; adjusting the emergency indication strategy based on the vehicle data; Provide emergency guidance for emergency parking strips in tunnels based on emergency indication strategies; Identify risky vehicles in tunnels based on vehicle data; Perform anomaly detection on risky vehicles and identify abnormal vehicles; Acquire the position coordinates of the abnormal vehicle; generate an emergency induction strategy for the abnormal vehicle based on the position coordinates, and perform emergency stop induction on the abnormal vehicle according to the emergency induction strategy.

Images